https://inferencebench.io/_next/static/chunks/3b1f6v-qd53j0.js

1(globalThis.TURBOPACK||(globalThis.TURBOPACK=[])).push(["object"==typeof document?document.currentScript:void 0,583517,e=>{"use strict";e.i(250825);var i=e.i(681307);let t=i.z.enum(["bf16","fp16","fp8","int8","int4","awq","gptq"]),a=i.z.enum(["vllm","sglang","tensorrt-llm","tgi","ollama","llama-cpp","exllama"]),_=i.z.object({type:i.z.enum(["dense","moe","hybrid"]),total_params_b:i.z.number().positive(),active_params_b:i.z.number().positive(),num_layers:i.z.number().int().positive(),hidden_dim:i.z.number().int().positive(),num_heads:i.z.number().int().positive(),num_kv_heads:i.z.number().int().positive(),head_dim:i.z.number().int().positive(),intermediate_dim:i.z.number().int().positive(),vocab_size:i.z.number().int().positive(),num_experts:i.z.number().int().positive().optional(),experts_per_token:i.z.number().int().positive().optional(),shared_experts:i.z.number().int().nonnegative().optional()}).strict(),s=i.z.object({weights_bf16_gb:i.z.number().positive(),weights_fp8_gb:i.z.number().positive(),weights_int4_gb:i.z.number().positive(),kv_per_token_bytes:i.z.number().nonnegative(),activation_estimate_gb:i.z.number().nonnegative()}).strict(),r=i.z.object({tools:i.z.boolean(),vision:i.z.boolean(),structured_output:i.z.boolean(),code:i.z.boolean(),math:i.z.boolean(),multilingual:i.z.boolean(),reasoning:i.z.boolean()}).strict(),n=i.z.object({input_per_m:i.z.number().nonnegative(),output_per_m:i.z.number().nonnegative(),reasoning_per_m:i.z.number().nonnegative().optional()}).strict(),o=i.z.string().min(1),d=i.z.object({type:i.z.enum(["dense","moe","gan","diffusion","encoder","encoder-decoder","pipeline","vocoder"]),total_params_b:i.z.number().positive().optional(),active_params_b:i.z.number().positive().optional(),num_layers:i.z.number().int().positive().optional(),hidden_dim:i.z.number().int().positive().optional(),notes:i.z.string().min(1)}).strict(),p=i.z.object({weights_bf16_gb:i.z.number().positive(),weights_fp8_gb:i.z.number().positive().optional(),weights_int4_gb:i.z.number().positive().optional(),activation_estimate_gb:i.z.number().nonnegative()}).strict(),m=i.z.discriminatedUnion("kind",[i.z.object({kind:i.z.literal("per_audio_second"),price:i.z.number().nonnegative()}),i.z.object({kind:i.z.literal("per_image"),price:i.z.number().nonnegative(),resolution:i.z.string().optional()}),i.z.object({kind:i.z.literal("per_video_minute"),price:i.z.number().nonnegative()}),i.z.object({kind:i.z.literal("per_embedding"),input_per_m_chars:i.z.number().nonnegative().optional(),output_per_m_embeddings:i.z.number().nonnegative().optional()}),i.z.object({kind:i.z.literal("per_inference"),price:i.z.number().nonnegative()}),i.z.object({kind:i.z.literal("free_open_weights")}),i.z.object({kind:i.z.literal("custom"),unit:i.z.string().min(1),price:i.z.number().nonnegative()})]),l=i.z.enum(["draft","active","deprecated","archived"]).default("active"),u=i.z.object({id:i.z.string().min(1),name:i.z.string().min(1),provider:i.z.string().min(1),family:i.z.string().min(1),status:l.optional(),version:i.z.string().optional(),release_date:i.z.string().regex(/^\d{4}-\d{2}-\d{2}$/),release_date_source:i.z.url().nullable().optional(),release_date_verified_at:i.z.iso.datetime().nullable().optional(),last_verified_at:i.z.iso.datetime().nullable().optional(),is_partial:i.z.boolean().optional(),deployment:i.z.enum(["self_hosted","api"]).optional(),modality:o.optional(),architecture:_.optional(),context_window:i.z.number().int().positive().optional(),supported_precisions:i.z.array(t).min(1).optional(),default_precision:t.optional(),memory:s.optional(),capabilities:r.optional(),supported_frameworks:i.z.array(a).optional(),api_pricing:i.z.record(i.z.string(),n).optional(),specialized_architecture:d.optional(),specialized_memory:p.optional(),specialized_pricing:m.optional(),description:i.z.string().optional(),intended_use:i.z.string().optional(),tags:i.z.array(i.z.string()).optional(),license:i.z.string().nullable().optional(),languages:i.z.array(i.z.string()).optional(),gated:i.z.boolean().optional(),library_name:i.z.string().optional(),huggingface_id:i.z.string().min(1).optional(),huggingface_id_status:i.z.enum(["verified","auto_resolved","manual","not_found","n_a"]).optional()}).strict().superRefine((e,i)=>{if(!0===e.is_partial){if("api"===e.deployment)for(let[t,a]of[["api_pricing",e.api_pricing],["capabilities",e.capabilities],["context_window",e.context_window]])null==a&&i.addIssue({code:"custom",path:[t],message:`${t} is required when deployment is "api"`});return}let t=e.modality??"text";for(let[a,_]of"text"===t?[["version",e.version],["architecture",e.architecture],["context_window",e.context_window],["supported_precisions",e.supported_precisions

1],["default_precision",e.default_precision],["memory",e.memory],["capabilities",e.capabilities],["supported_frameworks",e.supported_frameworks],["api_pricing",e.api_pricing]]:[["specialized_architecture",e.specialized_architecture],["specialized_pricing",e.specialized_pricing],["description",e.description]])null==_&&i.addIssue({code:"custom",path:[a],message:`${a} is required when is_partial is not true (modality=${t})`})});function g(e){return void 0===e.modality||"text"===e.modality}let c=[{id:"01-ai/yi-1.5-34b",name:"Yi 1.5 34B",provider:"01.AI",family:"Yi 1.5",version:"1.5",release_date:"2024-05-13",architecture:{type:"dense",total_params_b:34.4,active_params_b:34.4,num_layers:60,hidden_dim:7168,num_heads:56,num_kv_heads:8,head_dim:128,intermediate_dim:20480,vocab_size:64e3},context_window:2e5,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:68.8,weights_fp8_gb:34.4,weights_int4_gb:17.2,kv_per_token_bytes:245760,activation_estimate_gb:2},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm"],api_pricing:{"01ai":{input_per_m:.8,output_per_m:.8}},tags:["chat","open-weight","instruction-following"],huggingface_id:"01-ai/Yi-1.5-34B-Chat",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/01-ai/Yi-1.5-34B-Chat",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.176Z",license:"apache-2.0"},{id:"01-ai/yi-1.5-9b",name:"Yi 1.5 9B",provider:"01.AI",family:"Yi 1.5",version:"1.5",release_date:"2024-05-13",architecture:{type:"dense",total_params_b:8.83,active_params_b:8.83,num_layers:48,hidden_dim:4096,num_heads:32,num_kv_heads:4,head_dim:128,intermediate_dim:11008,vocab_size:64e3},context_window:4096,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:17.66,weights_fp8_gb:8.83,weights_int4_gb:4.415,kv_per_token_bytes:98304,activation_estimate_gb:1},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{"01ai":{input_per_m:.2,output_per_m:.2}},tags:["chat","open-weight","instruction-following"],huggingface_id:"01-ai/Yi-1.5-9B-Chat",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/01-ai/Yi-1.5-9B-Chat",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.177Z",license:"apache-2.0"},{id:"01-ai/yi-coder-9b",name:"Yi Coder 9B",provider:"01.AI",family:"Yi Coder",version:"1",release_date:"2024-09-04",architecture:{type:"dense",total_params_b:8.8,active_params_b:8.8,num_layers:48,hidden_dim:3584,num_heads:28,num_kv_heads:4,head_dim:128,intermediate_dim:18944,vocab_size:64e3},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:17.6,weights_fp8_gb:8.8,weights_int4_gb:4.4,kv_per_token_bytes:98304,activation_estimate_gb:1},capabilities:{tools:!1,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{},tags:["code","open-weight"],huggingface_id:"01-ai/Yi-Coder-9B-Chat",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/01-ai/Yi-Coder-9B-Chat",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.177Z",license:"apache-2.0"},{id:"01-ai/yi-large",name:"Yi-Large",provider:"01.AI",family:"Yi",version:"1.0",release_date:"2024-05-13",architecture:{type:"moe",total_params_b:102.6,active_params_b:24,num_layers:64,hidden_dim:8192,num_heads:64,num_kv_heads:8,head_dim:128,intermediate_dim:24576,vocab_size:64e3,num_experts:32,experts_per_token:4},context_window:32768,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:205.2,weights_fp8_gb:102.6,weights_int4_gb:51.3,kv_per_token_bytes:262144,activation_estimate_gb:2.5},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang"],api_pricing:{"01ai":{input_per_m:3,output_per_m:3}},tags:["chat","moe","estimated-architecture","api-only"],release_date_source:"https://platform.01.ai/docs/api-reference/models",license:"proprietary",release_date_verified_at:"2026-06-25T16:20:15.665Z",last_verified_at:"2026-06-25T16:20:15.665Z"},{id:"01-ai/yi-lightning",name:"Yi-Lightning",provider:"01.AI",family:"Yi",version:"Lightning",release_date:"2024-10-01",architecture:{type:"moe",total_params_b:200,active_params_b:22,num_layers:64,hidden_dim:6144,num_heads:48,num_kv_heads:8,head_dim:128,intermediate_dim:16384,vocab_size:64e3,num_experts:32,experts_per_token:4},context_window:16384,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:400,weights_fp8_gb:200,weights_int4_gb:100,kv_per_token_bytes:131072,activation_estimate_gb:2},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm","sglang"],api_pricing:{"01ai":{input_per_m:.99,output_per_m:.99}},tags:["chat","code","reasoning","moe","estimated-architecture","api-only"],release_date_source:"https://platform.01.ai/docs/api-reference/models",license:"proprietary",release_date_verified_at:"2026-06-25T16:20:15.668Z",last_verified_at:"2026-06-25T16:20:15.668Z"}

1,{id:"Qwen/Qwen3-235B-A22B-Thinking-2507",name:"Qwen3-235B-A22B-Thinking-2507",provider:"Qwen",family:"Qwen3",version:"3.0",release_date:"2025-07-21",architecture:{type:"moe",total_params_b:235,active_params_b:22,num_layers:94,hidden_dim:4096,num_heads:64,num_kv_heads:4,head_dim:128,intermediate_dim:12288,vocab_size:151936,num_experts:128,experts_per_token:8},context_window:262144,supported_precisions:["bf16"],default_precision:"bf16",memory:{weights_bf16_gb:470,weights_fp8_gb:235,weights_int4_gb:117.5,kv_per_token_bytes:192512,activation_estimate_gb:.001540096},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm","sglang"],api_pricing:{openrouter:{input_per_m:.13,output_per_m:.6}},tags:["chat","code","reasoning","math","multilingual"],status:"active",huggingface_id:"Qwen/Qwen3-235B-A22B-Thinking-2507",huggingface_id_status:"manual",release_date_source:"https://huggingface.co/Qwen/Qwen3-235B-A22B-Thinking-2507",last_verified_at:"2026-06-25T16:19:04.531Z",license:"apache-2.0"},{id:"ai21/jamba-1.5-large",name:"Jamba 1.5 Large",provider:"AI21",family:"Jamba",version:"1.5",release_date:"2024-08-22",architecture:{type:"hybrid",total_params_b:398,active_params_b:52,num_layers:64,hidden_dim:8192,num_heads:64,num_kv_heads:8,head_dim:128,intermediate_dim:24576,vocab_size:65536,num_experts:16,experts_per_token:2},context_window:256e3,supported_precisions:["bf16","fp8"],default_precision:"bf16",memory:{weights_bf16_gb:796,weights_fp8_gb:398,weights_int4_gb:199,kv_per_token_bytes:131072,activation_estimate_gb:3},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang"],api_pricing:{ai21:{input_per_m:2,output_per_m:8}},tags:["chat","moe","ssm-hybrid","open-weight","long-context"],huggingface_id:"ai21labs/AI21-Jamba-Large-1.5",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/ai21labs/AI21-Jamba-Large-1.5",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.177Z",license:"other"},{id:"ai21/jamba-1.5-mini",name:"Jamba 1.5 Mini",provider:"AI21",family:"Jamba",version:"1.5",release_date:"2024-08-22",architecture:{type:"hybrid",total_params_b:52,active_params_b:12,num_layers:32,hidden_dim:4096,num_heads:32,num_kv_heads:8,head_dim:128,intermediate_dim:14336,vocab_size:65536,num_experts:16,experts_per_token:2},context_window:256e3,supported_precisions:["bf16","fp8"],default_precision:"bf16",memory:{weights_bf16_gb:104,weights_fp8_gb:52,weights_int4_gb:26,kv_per_token_bytes:65536,activation_estimate_gb:1.5},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang"],api_pricing:{ai21:{input_per_m:.2,output_per_m:.4}},tags:["chat","moe","ssm-hybrid","open-weight","long-context"],huggingface_id:"ai21labs/AI21-Jamba-Mini-1.5",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/ai21labs/AI21-Jamba-Mini-1.5",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.177Z",license:"other"},{id:"alibaba/gte-qwen2-7b",name:"GTE Qwen2 7B",provider:"Alibaba",family:"GTE",version:"2",release_date:"2024-06-17",architecture:{type:"dense",total_params_b:7.6,active_params_b:7.6,num_layers:32,hidden_dim:3584,num_heads:28,num_kv_heads:4,head_dim:128,intermediate_dim:18944,vocab_size:152064},context_window:32768,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:15.2,weights_fp8_gb:7.6,weights_int4_gb:3.8,kv_per_token_bytes:65536,activation_estimate_gb:.8},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm"],api_pricing:{together:{input_per_m:.016,output_per_m:.016}},tags:["embedding","open-weight"],huggingface_id:"Alibaba-NLP/gte-Qwen2-7B-instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/Alibaba-NLP/gte-Qwen2-7B-instru

1ct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.177Z",license:"apache-2.0"},{id:"alibaba/marco-o1",name:"Marco O1",provider:"Alibaba",family:"Marco",version:"O1",release_date:"2024-11-21",architecture:{type:"dense",total_params_b:7.6,active_params_b:7.6,num_layers:28,hidden_dim:3584,num_heads:28,num_kv_heads:4,head_dim:128,intermediate_dim:18944,vocab_size:152064},context_window:65536,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:15.2,weights_fp8_gb:7.6,weights_int4_gb:3.8,kv_per_token_bytes:57344,activation_estimate_gb:.8},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm","sglang","tgi","ollama"],api_pricing:{},tags:["chat","multilingual","reasoning","open-weight"],huggingface_id:"AIDC-AI/Marco-o1",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/AIDC-AI/Marco-o1",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.177Z",license:"apache-2.0"},{id:"alibaba/qwen-1.5-moe-a2.7b",name:"Qwen 1.5 MoE A2.7B",provider:"Alibaba",family:"Qwen 1.5",version:"1.5-moe",release_date:"2024-03-28",architecture:{type:"moe",total_params_b:14.3,active_params_b:2.7,num_layers:24,hidden_dim:2048,num_heads:16,num_kv_heads:16,head_dim:128,intermediate_dim:5632,vocab_size:151936,num_experts:60,experts_per_token:4},context_window:32768,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:28.6,weights_fp8_gb:14.3,weights_int4_gb:7.15,kv_per_token_bytes:196608,activation_estimate_gb:.5},capabilities:{tools:!1,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi"],api_pricing:{},tags:["chat","code","moe","open-weight"],huggingface_id:"Qwen/Qwen1.5-MoE-A2.7B",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/Qwen/Qwen1.5-MoE-A2.7B",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.177Z",license:"other"},{id:"alibaba/qwen-2-audio-7b",name:"Qwen 2 Audio 7B",provider:"Alibaba",family:"Qwen 2",version:"2-audio",release_date:"2024-07-15",architecture:{type:"dense",total_params_b:7.6,active_params_b:7.6,num_layers:28,hidden_dim:3584,num_heads:28,num_kv_heads:4,head_dim:128,intermediate_dim:18944,vocab_size:152064},context_window:32768,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:15.2,weights_fp8_gb:7.6,weights_int4_gb:3.8,kv_per_token_bytes:57344,activation_estimate_gb:1},capabilities:{tools:!1,vision:!1,structured_output:!0,code:!1,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi"],api_pricing:{},tags:["audio","multimodal","open-weight"],huggingface_id:"Qwen/Qwen2-Audio-7B-Instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/Qwen/Qwen2-Audio-7B-Instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.177Z",license:"apache-2.0"},{id:"alibaba/qwen-2.5-14b",name:"Qwen 2.5 14B",provider:"Alibaba",family:"Qwen 2.5",version:"2.5",release_date:"2024-09-19",architecture:{type:"dense",total_params_b:14.8,active_params_b:14.8,num_layers:48,hidden_dim:5120,num_heads:40,num_kv_heads:8,head_dim:128,intermediate_dim:13824,vocab_size:152064},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:29.6,weights_fp8_gb:14.8,weights_int4_gb:7.4,kv_per_token_bytes:196608,activation_estimate_gb:1.5},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm"],api_pricing:{alibaba:{input_per_m:.4,output_per_m:.4},together:{input_per_m:.3,output_per_m:.3}},tags:["chat","open-weight","instruction-following"],huggingface_id:"Qwen/Qwen2.5-Coder-14B-Instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/Qwen/Qwen2.5-Coder-14B-Instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.177Z",license:"apache-2.0"},{id:"alibaba/qwen-2.5-32b",name:"Qwen 2.5 32B",provider:"Alibaba",family:"Qwen 2.5",version:"2.5",release_date:"2024-09-19",architecture:{type:"dense",total_params_b:32.5,active_params_b:32.5,num_layers:64,hidden_dim:5120,num_heads:40,num_kv_heads:8,head_dim:128,intermediate_dim:27648,vocab_size:152064},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:65,weights_fp8_gb:32.5,weights_int4_gb:16.25,kv_per_token_bytes:262144,activation_estimate_gb:2},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm"],api_pricing:{alibaba:{input_per_m:.8,output_per_m:.8},together:{input_per_m:.8,output_per_m:.8}},tags:["chat","open-weight","instruction-following"],huggingface_id:"Qwen

1/Qwen2.5-32B-Instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/Qwen/Qwen2.5-32B-Instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.177Z",license:"apache-2.0"},{id:"alibaba/qwen-2.5-3b",name:"Qwen 2.5 3B",provider:"Alibaba",family:"Qwen 2.5",version:"2.5",release_date:"2024-09-19",architecture:{type:"dense",total_params_b:3.09,active_params_b:3.09,num_layers:36,hidden_dim:2048,num_heads:16,num_kv_heads:2,head_dim:128,intermediate_dim:11008,vocab_size:152064},context_window:32768,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:6.18,weights_fp8_gb:3.09,weights_int4_gb:1.545,kv_per_token_bytes:36864,activation_estimate_gb:.5},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","ollama"],api_pricing:{alibaba:{input_per_m:.1,output_per_m:.1}},tags:["chat","open-weight","instruction-following"],huggingface_id:"Qwen/Qwen2.5-3B-Instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/Qwen/Qwen2.5-3B-Instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.178Z",license:"other"}

1,{id:"alibaba/qwen-2.5-coder-32b",name:"Qwen 2.5 Coder 32B",provider:"Alibaba",family:"Qwen 2.5",version:"2.5",release_date:"2024-11-12",architecture:{type:"dense",total_params_b:32.5,active_params_b:32.5,num_layers:64,hidden_dim:5120,num_heads:40,num_kv_heads:8,head_dim:128,intermediate_dim:27648,vocab_size:152064},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:65,weights_fp8_gb:32.5,weights_int4_gb:16.25,kv_per_token_bytes:262144,activation_estimate_gb:2},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm"],api_pricing:{alibaba:{input_per_m:.8,output_per_m:.8},together:{input_per_m:.8,output_per_m:.8},featherless:{input_per_m:0,output_per_m:0},scaleway:{input_per_m:0,output_per_m:0},openrouter:{input_per_m:.66,output_per_m:1}},tags:["code","open-weight","code-specialized"],huggingface_id:"Qwen/Qwen2.5-Coder-32B-Instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/Qwen/Qwen2.5-Coder-32B-Instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.178Z",license:"apache-2.0"},{id:"allenai/molmo-72b",name:"Molmo 72B",provider:"Allen AI",family:"Molmo",version:"1",release_date:"2024-09-25",architecture:{type:"dense",total_params_b:72,active_params_b:72,num_layers:80,hidden_dim:8192,num_heads:64,num_kv_heads:8,head_dim:128,intermediate_dim:28672,vocab_size:152064},context_window:8192,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:144,weights_fp8_gb:72,weights_int4_gb:36,kv_per_token_bytes:589824,activation_estimate_gb:3},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm","sglang"],api_pricing:{},tags:["multimodal","vision","open-weight"],huggingface_id:"allenai/Molmo-72B-0924",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/allenai/Molmo-72B-0924",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.178Z",license:"apache-2.0"},{id:"allenai/olmo-2-13b",name:"OLMo 2 13B",provider:"Allen AI",family:"OLMo 2",version:"2",release_date:"2024-12-19",architecture:{type:"dense",total_params_b:13,active_params_b:13,num_layers:40,hidden_dim:5120,num_heads:40,num_kv_heads:40,head_dim:128,intermediate_dim:13824,vocab_size:100278},context_window:4096,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:26,weights_fp8_gb:13,weights_int4_gb:6.5,kv_per_token_bytes:819200,activation_estimate_gb:1.5},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!0,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","ollama"],api_pricing:{},tags:["chat","open-weight","open-source"],huggingface_id:"allenai/OLMo-2-1124-13B",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/allenai/OLMo-2-1124-13B",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.178Z",license:"apache-2.0"},{id:"allenai/olmo-2-7b",name:"OLMo 2 7B",provider:"Allen AI",family:"OLMo 2",version:"2",release_date:"2024-12-19",architecture:{type:"dense",total_params_b:7,active_params_b:7,num_layers:32,hidden_dim:4096,num_heads:32,num_kv_heads:32,head_dim:128,intermediate_dim:11008,vocab_size:100278},context_window:4096,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:14,weights_fp8_gb:7,weights_int4_gb:3.5,kv_per_token_bytes:524288,activation_estimate_gb:1},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!0,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","ollama"],api_pricing:{},tags:["chat","open-weight","open-source"],huggingface_id:"allenai/OLMo-2-1124-7B",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/allenai/OLMo-2-1124-7B",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.178Z",license:"apache-2.0"},{id:"allenai/qwen35-9b-cli-gym",name:"qwen35-9b-cli-gym",gated:!1,family:"allenai",provider:"allenai",capabilities:{code:!1,math:!1,tools:!0,vision:!1,reasoning:!1,multilingual:!1,structured_output:!1},last_verified_at:"2026-06-23T06:56:11.414Z",default_precision:"bf16",release_date_source:"https://huggingface.co/allenai/qwen35-9b-cli-gym",supported_precisions:["bf16","fp8","int4"],status:"draft",is_partial:!0,release_date:"2026-06-23"},{id:"amazon/nova-lite",name:"Amazon Nova Lite",provider:"Amazon",family:"Nova",version:"1",release_date:"2024-12-03",architecture:{type:"dense",total_params_b:12,active_params_b:12,num_layers:40,hidden_dim:5120,num_heads:40,num_kv_heads:8,head_dim:128,intermediate_dim:14336,vocab_size:128e3},context_window:3e5,supported_precisions:["bf16","fp8"],default_precision:"bf16",memory:{weights_bf16_gb:24,weights_fp8_gb:12,weights_int4_gb:6,kv_per_token_bytes:163840,activation_estimate_gb:1},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm"],api_pricing:{amazon:{input_per_m:.06,output_per_m:.24},openrouter:{input_per_m:.06,output_per_m:.24}},tags:["chat","multimodal","estimated-architecture","api-only"],huggingface_id_status:"n_a",release_date_source:"https://docs.aws.amazon.com/nova/latest/userguide/what-is-nova.html",license:"proprietary",release_date_verified_at:"2026-06-25T16:20:15.668Z",last_verified_at:"2026-06-25T16:20:15.668Z"},{id:"amazon/nova-pro",name:"Amazon Nova Pro",provider:"Amazon",family:"Nova",version:"1",release_date:"2024-12-03",architecture:{type:"dense",total_params_b:50,active_params_b:50,num_layers:60,hidden_dim:8192,num_heads:64,num_kv_heads:8,head_dim:128,intermediate_dim:28672,vocab_size:128e3},context_window:3e5,supported_precisions:["bf16","fp8"],default_precision:"bf16",memory:{weights_bf16_gb:100,weights_fp8_gb:50,weights_int4_gb:25,kv_per_token_bytes:245760,activation_estimate_gb:2},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm"],api_pricing:{amazon:{input_per_m:.8,output_per_m:3.2},openrouter:{input_per_m:.7999999999999999,output_per_m:3.1999999999999997}},tags:["chat","multimodal","estimated-architecture","api-only"],huggingface_id_status:"n_a",release_date_source:"https://docs.aws.amazon.com/nova/latest/userguide/what-is-nova.html",license:"proprietary",release_date_verified_at:"2026-06-25T16:20:15.668Z",last_verified_at:"2026-06-25T16:20:15.668Z"},{id:"answerdotai/qwen2.5-32b-instruct-cla2-adj-fp8kv-gradual-unfreeze",name:"Qwen2.5-32B-Instruct-CLA2-adj-fp8KV-gradual-unfreeze",family:"answerdotai",license:null,version:"2.5",provider:"answerdotai",is_partial:!0,capabilities:{code:!1,math:!1,tools:!1,vision:!1,reasoning:!1,multilingual:!1,structured_output:!1},release_date:"2024-10-29",last_verified_at:"2026-06-22T11:43:53.950Z",release_date_source:"https://huggingface.co/answerdotai/Qwen2.5-32B-Instruct-CLA2-adj-fp8KV-gradual-unfreeze",release_date_verified_at:"2026-06-22T11:43:53.950Z",status:"draft"},{id:"proprietary/claude-3-opus",name:"Claude 3 Opus",provider:"Anthropic",family:"Claude",version:"3-opus",release_date:"2024-03-04",architecture:{type:"dense",total_params_b:175,active_params_b:175,num_layers:80,hidden_dim:12288,num_heads:96,num_kv_heads:16,head_dim:128,intermediate_dim:49152,vocab_size:152064},context_window:2e5,supported_precisions:["bf16"],default_precision:"bf16",memory:{weights_bf16_gb:350,weights_fp8_gb:175,weights_int4_gb:87.5,kv_per_token_bytes:327680,activation_estimate_gb:4.5},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:[],api_pricing:{anthropic:{input_per_m:15,output_per_m:75}},tags:["proprietary","api-only","estimated-architecture","legacy"],huggingface_id_status:"n_a",release_date_source:"https://docs.anthropic.com/en/docs/about-claude/models",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-16T09:29:53.000Z"},{id:"proprietary/claude-3-sonnet",name:"Claude 3 Sonnet",provider:"Anthropic",family:"Claude",version:"3-sonnet",release_date:"2024-03-04",architecture:{type:"dense",total_params_b:70,active_params_b:70,num_layers:64,hidden_dim:8192,num_heads:64,num_kv_heads:8,head_dim:128,intermediate_dim:28672,vocab_size:152064},context_window:2e5,supported_precisions:["bf16"],default_precision:"bf16",memory:{weights_bf16_gb:140,weights_fp8_gb:70,weights_int4_gb:35,kv_per_token_bytes:131072,activation_estimate_gb:3},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:[],api_pricing:{anthropic:{input_per_m:3,output_per_m:15}},tags:["proprietary","api-only","estimated-architecture","legacy"],huggingface_id_status:"n_a",release_date_source:"https://docs.anthropic.com/en/docs/about-claude/models",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-16T09:29:53.000Z"},{id:"anthropic/claude-fable-5",name:"Claude Fable 5",family:"Claude",provider:"Anthropic",is_partial:!0,release_date:"2026-06-09",release_date_source:"https://openrouter.ai/models/anthropic/claude-fable-5",release_date_verified_at:"2026-06-20T16:00:00Z",last_verified_at:"2026-06-20T16:00:00Z",context_window:1e6,deployment:"api",api_pricing:{anthropic:{input_per_m:10,output_per_m:50}},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},description:"Claude Fable 5 is Anthropic's high-capability Claude 5-family model — multimodal, tool-using and reasoning-capable with a 1M-token context window, served through the Claude API."},{id:"proprietary/claude-haiku-3.5",name:"Claude 3.5 Haiku",provider:"Anthropic",family:"Claude",version:"3.5-haiku",release_date:"2024-10-29",architecture:{type:"dense",total_params_b:20,active_params_b:20,num_layers:40,hidden_dim:5120,num_heads:40,num_kv_heads:8,head_dim:128,intermediate_dim:13824,vocab_size:152064},context_window:2e5,supported_precisions:["bf16"],default_precision:"bf16",memory:{weights_bf16_gb:40,weights_fp8_gb:20,weights_int4_gb:10,kv_per_token_bytes:40960,activation_estimate_gb:1.5},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:[],api_pricing:{anthropic:{input_per_m:.8,output_per_m:4},openrouter:{input_per_m:.7999999999999999,output_per_m:4}},tags:["proprietary","api-only","estimated-architecture","fast"],huggingface_id_status:"n_a",release_date_source:"https://docs.anthropic.com/en/docs/about-claude/models",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-16T09:29:53.000Z"},{id:"proprietary/claude-haiku-4-5",name:"Claude Haiku 4.5",provider:"Anthropic",family:"Claude",version:"4.5",release_date:"2025-10-01",release_date_source:"https://docs.anthropic.com/en/docs/about-claude/models/overview",release_date_verified_at:"2026-05-16T14:20:26Z",last_verified_at:"2026-05-16T14:20:26Z",is_partial:!1,tags:["fast","small"],

1context_window:2e5,supported_precisions:["bf16"],default_precision:"bf16",architecture:{type:"moe",total_params_b:30,active_params_b:30,num_layers:40,hidden_dim:8192,num_heads:64,num_kv_heads:8,head_dim:128,intermediate_dim:36864,vocab_size:2e5,num_experts:16,experts_per_token:2},memory:{weights_bf16_gb:60,weights_fp8_gb:30,weights_int4_gb:15,kv_per_token_bytes:204800,activation_estimate_gb:4},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm"],api_pricing:{anthropic:{input_per_m:1,output_per_m:5}}},{id:"proprietary/claude-opus-4-1",name:"Claude Opus 4.1",provider:"Anthropic",family:"Claude",version:"4.1",release_date:"2025-08-05",release_date_source:"https://docs.anthropic.com/en/docs/about-claude/models/overview",release_date_verified_at:"2026-05-16T14:20:26Z",last_verified_at:"2026-05-16T14:20:26Z",is_partial:!1,tags:["frontier"],context_window:2e5,supported_precisions:["bf16"],default_precision:"bf16",architecture:{type:"moe",total_params_b:300,active_params_b:75,num_layers:96,hidden_dim:12288,num_heads:96,num_kv_heads:10,head_dim:128,intermediate_dim:36864,vocab_size:2e5,num_experts:16,experts_per_token:2},memory:{weights_bf16_gb:600,weights_fp8_gb:300,weights_int4_gb:150,kv_per_token_bytes:204800,activation_estimate_gb:4},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm"],api_pricing:{anthropic:{input_per_m:15,output_per_m:75}}},{id:"proprietary/claude-opus-4-5",name:"Claude Opus 4.5",provider:"Anthropic",family:"Claude",version:"4.5",release_date:"2025-11-01",release_date_source:"https://docs.anthropic.com/en/docs/about-claude/models/overview",release_date_verified_at:"2026-05-16T14:20:26Z",last_verified_at:"2026-05-16T14:20:26Z",is_partial:!1,tags:["frontier"],context_window:2e5,supported_precisions:["bf16"],default_precision:"bf16",architecture:{type:"moe",total_params_b:400,active_params_b:80,num_layers:96,hidden_dim:12288,num_heads:96,num_kv_heads:10,head_dim:128,intermediate_dim:36864,vocab_size:2e5,num_experts:16,experts_per_token:2},memory:{weights_bf16_gb:800,weights_fp8_gb:400,weights_int4_gb:200,kv_per_token_bytes:204800,activation_estimate_gb:4},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm"],api_pricing:{anthropic:{input_per_m:5,output_per_m:25}}},{name:"Claude Opus 4.6",provider:"Anthropic",family:"Claude",version:"4.6",release_date:"2025-12-15",release_date_source:"https://docs.anthropic.com/en/docs/about-claude/models/overview",release_date_verified_at:"2026-05-16T14:43:00Z",last_verified_at:"2026-05-16T14:43:00Z",context_window:1e6,supported_precisions:["bf16"],default_precision:"bf16",architecture:{type:"moe",total_params_b:450,active_params_b:85,num_layers:80,hidden_dim:10240,num_heads:80,num_kv_heads:10,head_dim:128,intermediate_dim:36864,vocab_size:2e5,num_experts:16,experts_per_token:2},memory:{weights_bf16_gb:900,weights_fp8_gb:450,weights_int4_gb:225,kv_per_token_bytes:204800,activation_estimate_gb:4},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm"],api_pricing:{anthropic:{input_per_m:5,output_per_m:25}},is_partial:!1,id:"proprietary/claude-opus-4-6"},{name:"Claude Opus 4.7",provider:"Anthropic",family:"Claude",version:"4.7",release_date:"2026-01-15",release_date_source:"https://docs.anthropic.com/en/docs/about-claude/models/overview",release_date_verified_at:"2026-05-16T14:43:00Z",last_verified_at:"2026-05-16T14:43:00Z",context_window:1e6,supported_precisions:["bf16"],default_precision:"bf16",architecture:{type:"moe",total_params_b:500,active_params_b:90,num_layers:80,hidden_dim:10240,num_heads:80,num_kv_heads:10,head_dim:128,intermediate_dim:36864,vocab_size:2e5,num_experts:16,experts_per_token:2},memory:{weights_bf16_gb:1e3,weights_fp8_gb:500,weights_int4_gb:250,kv_per_token_bytes:204800,activation_estimate_gb:4},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm"],api_pricing:{anthropic:{input_per_m:5,output_per_m:25}},is_partial:!1,id:"proprietary/claude-opus-4-7"},{id:"anthropic/claude-opus-4.8",name:"Claude Opus 4.8",family:"Claude",provider:"Anthropic",is_partial:!0,release_date:"2026-05-27",release_date_source:"https://openrouter.ai/models/anthropic/claude-opus-4.8",release_date_verified_at:"2026-06-20T16:00:00Z",last_verified_at:"2026-06-20T16:00:00Z",context_window:1e6,deployment:"api",api_pricing:{anthropic:{input_per_m:5,output_per_m:25}},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},description:"Claude Opus 4.8 is Anthropic's flagship model — a frontier, multimodal, tool-using and reasoning model with a 1M-token context window, available through the Claude API."},{id:"proprietary/claude-opus-4",name:"Claude Opus 4",provider:"Anthropic",family:"Claude",version:"opus-4",release_date:"2025-05-22",release_date_source:"https://www.anthropic.com/news/claude-4",release_date_verified_at:"2026-05-13T14:42:10Z",last_verified_at:"2026-05-13T14:42:10Z",architecture:{type:"dense",total_params_b:200,active_params_b:200,num_layers:96,hidden_dim:12288,num_heads:96,num_kv_heads:16,head_dim:128,intermediate_dim:49152,vocab_size:152064},context_window:2e5,supported_precisions:["bf16"],default_precision:"bf16",memory:{weights_bf16_gb:400,weights_fp8_gb:200,weights_int4_gb:100,kv_per_token_bytes:393216,activation_estimate_gb:5},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:[],api_pricing:{anthropic:{input_per_m:15,output_per_m:75},openrouter:{input_per_m:15,output_per_m:75}},tags:["proprietary","api-only","estimated-architecture","flagship","reasoning"],huggingface_id_status:"n_a"},{name:"Claude 3.5 Sonnet",provider:"Anthropic",family:"Claude",version:"3.5",release_date:"2024-06-20",release_date_source:"https://www.anthropic.com/news/claude-3-5-sonnet",release_date_verified_at:"2026-05-18T12:00:00Z",last_verified_at:"2026-05-18T12:00:00Z",context_window:2e5,supported_precisions:["bf16"],default_precision:"bf16",architecture:{type:"moe",total_params_b:175,active_params_b:50,num_layers:60,hidden_dim:8192,num_heads:64,num_kv_heads:8,head_dim:128,intermediate_dim:28672,vocab_size:2e5,num_experts:8,experts_per_token:2},memory:{weights_bf16_gb:350,weights_fp8_gb:175,weights_int4_gb:88,kv_per_token_bytes:163840,activation_estimate_gb:4},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm"],api_pricing:{anthropic:{input_per_m:3,output_per_m:15}},is_partial:!1,id:"proprietary/claude-sonnet-3-5"},{id:"proprietary/claude-sonnet-4-5",name:"Claude Sonnet 4.5",provider:"Anthropic",family:"Claude",version:"4.5",release_date:"2025-09-29",release_date_source:"https://docs.anthropic.com/en/docs/about-claude/models/overview",release_date_verified_at:"2026-05-16T14:20:26Z",last_verified_at:"2026-05-16T14:20:26Z",is_partial:!1,tags:["mid-tier"],context_window:2e5,supported_precisions:["bf16"],default_precision:"bf16",architecture:{type:"moe",total_params_b:150,active_params_b:60,num_layers:64,hidden_dim:10240,num_heads:80,num_kv_heads:10,head_dim:128,intermediate_dim:36864,vocab_size:2e5,num_experts:16,experts_per_token:2},memory:{weights_bf16_gb:300,weights_fp8_gb:150,weights_int4_gb:75,kv_per_token_bytes:204800,activation_estimate_gb:4},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm"],api_pricing:{anthropic:{input_per_m:3,output_per_m:15}}},{name:"Claude Sonnet 4.6",provider:"Anthropic",family:"Claude",version:"4.6",release_date:"2026-01-15",release_date_source:"https://docs.anthropic.com/en/docs/about-claude/models/overview",release_date_verified_at:"2026-05-16T14:43:00Z",last_verified_at:"2026-05-16T14:43:00Z",context_window:1e6,supported_precisions:["bf16"],default_precision:"bf16",architecture:{type:"moe",total_params_b:180,active_params_b:70,num_layers:80,hidden_dim:10240,num_heads:80,num_kv_heads:10,head_dim:128,intermediate_dim:36864,vocab_size:2e5,num_experts:16,experts_per_token:2},memory:{weights_bf16_gb:360,weights_fp8_gb:180,weights_int4_gb:90,kv_per_token_bytes:204800,activation_estimate_gb:4},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm"],api_pricing:{anthropic:{input_per_m:3,output_per_m:15}},is_partial:!1,id:"proprietary/claude-sonnet-4-6"},{id:"proprietary/claude-sonnet-4",name:"Claude Sonnet 4",provider:"Anthropic",family:"Claude",version:"sonnet-4",release_date:"2025-05-22",release_date_source:"https://www.anthropic.com/news/claude-4",release_date_verified_at:"2026-05-13T14:42:10Z",last_verified_at:"2026-05-13T14:42:10Z",architecture:{type:"dense",total_params_b:70,active_params_b:70,num_layers:64,hidden_dim:8192,num_heads:64,num_kv_heads:8,head_dim:128,intermediate_dim:28672,vocab_size:152064},context_window:2e5,supported_precisions:["bf16"],default_precision:"bf16",memory:{weights_bf16_gb:140,weights_fp8_gb:70,weights_int4_gb:35,kv_per_token_bytes:131072,activation_estimate_gb:3},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:[],api_pricing:{anthropic:{input_per_m:3,output_per_m:15},openrouter:{input_per_m:3,output_per_m:15}},tags:["proprietary","api-only","estimated-architecture","flagship"],huggingface_id_status:"n_a"},{id:"apple/openelm-3b",name:"OpenELM 3B",provider:"Apple",family:"OpenELM",version:"1",release_date:"2024-04-22",architecture:{type:"dense",total_params_b:3,active_params_b:3,num_layers:36,hidden_dim:3072,num_heads:24,num_kv_heads:4,head_dim:128,intermediate_dim:12288,vocab_size:32e3},context_window:2048,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:6,weights_fp8_gb:3,weights_int4_gb:1.5,kv_per_token_bytes:73728,activation_estimate_gb:.3},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","ollama"],api_pricing:{},tags:["chat","open-weight","lightweight","mobile","estimated-architecture"],huggingface_id:"apple/OpenELM-3B-Instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/apple/OpenELM-3B-Instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.178Z",license:"apple-amlr"},{id:"arabic/jais-30b",name:"JAIS 30B",provider:"G42/Inception",family:"JAIS",version:"1",release_date:"2024-02-19",architecture:{type:"dense",total_params_b:30,active_params_b:30,num_layers:56,hidden_dim:7168,num_heads:56,num_kv_heads:56,head_dim:128,intermediate_dim:19456,vocab_size:84992},context_window:8192,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:60,weights_fp8_gb:30,weights_int4_gb:15,kv_per_token_bytes:802816,activation_estimate_gb:1.5},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","tgi"],api_pricing:{},tags:["arabic","multilingual","open-weight"],huggingface_id:"inceptionai/jais-30b-chat-v3",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/inceptionai/jais-30b-chat-v3",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-16T09:29:53.000Z"},{id:"baai/bge-base-en-v1.5",name:"BGE Base EN v1.5",provider:"BAAI",family:"BGE",version:"1.5",release_date:"2023-09-12",architecture:{type:"dense",total_params_b:.11,active_params_b:.11,num_layers:12,hidden_dim:768,num_heads:12,num_kv_heads:12,head_dim:64,intermediate_dim:3072,vocab_size:30522},context_window:512,supported_precisions:["bf16","fp8"],default_precision:"bf16",memory:{weights_bf16_gb:.22,weights_fp8_gb:.11,weights_int4_gb:.055,kv_per_token_bytes:18432,activation_estimate_gb:.05},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","tgi"],api_pricing:{},tags:["embedding","open-weight"],huggingface_id:"BAAI/bge-base-en-v1.5",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/BAAI/bge-base-en-v1.5",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.178Z",license:"mit"},{id:"baai/bge-large-en-v1.5",name:"BGE Large EN v1.5",provider:"BAAI",family:"BGE",version:"1.5",release_date:"2023-09-12",architecture:{type:"dense",total_params_b:.335,active_params_b:.335,num_layers:24,hidden_dim:1024,num_heads:16,num_kv_heads:16,head_dim:64,intermediate_dim:4096,vocab_size:30522},context_window:512,supported_precisions:["bf16","fp8"],default_precision:"bf16",memory:{weights_bf16_gb:.67,weights_fp8_gb:.335,weights_int4_gb:.168,kv_per_token_bytes:98304,activation_estimate_gb:.1},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","tgi","tensorrt-llm"],api_pricing:{together:{input_per_m:.008,output_per_m:.008}},tags:["embedding","open-weight"],huggingface_id:"BAAI/bge-large-en-v1.5",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/BAAI/bge-large-en-v1.5",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.178Z",license:"mit"},{id:"baai/bge-m3",name:"BGE M3",provider:"BAAI",family:"BGE",version:"M3",release_date:"2024-01-30",architecture:{type:"dense",total_params_b:.568,active_params_b:.568,num_layers:24,hidden_dim:1024,num_heads:16,num_kv_heads:16,head_dim:64,intermediate_dim:4096,vocab_size:250002},context_window:8192,supported_precisions:["bf16","fp8"],default_precision:"bf16",memory:{weights_bf16_gb:1.136,weights_fp8_gb:.568,weights_int4_gb:.284,kv_per_token_bytes:98304,activation_estimate_gb:.1},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","tgi","tensorrt-llm"],api_pricing:{together:{input_per_m:.008,output_per_m:.008}},tags:["embedding","multilingual","open-weight"],huggingface_id:"BAAI/bge-m3",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/BAAI/bge-m3",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.178Z",license:"mit"},{id:"baai/bge-small-en-v1.5",name:"BGE Small EN v1.5",provider:"BAAI",family:"BGE",version:"1.5",release_date:"2023-09-12",architecture:{type:"dense",total_params_b:.033,active_params_b:.033,num_layers:12,hidden_dim:384,num_heads:12,num_kv_heads:12,head_dim:32,intermediate_dim:1536,vocab_size:30522},context_window:512,supported_precisions:["bf16","fp8"],default_precision:"bf16",memory:{weights_bf16_gb:.066,weights_fp8_gb:.033,weights_int4_gb:.017,kv_per_token_bytes:9216,activation_estimate_gb:.05},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","tgi"],api_pricing:{},tags:["embedding","open-weight","lightweight"],huggingface_id:"BAAI/bge-small-en-v1.5",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/BAAI/bge-small-en-v1.5",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.178Z",license:"mit"},{id:"baai/robobrain-dex",name:"RoboBrain-Dex",family:"BAAI",license:"apache-2.0",provider:"BAAI",is_partial:!0,capabilities:{code:!1,math:!1,tools:!1,vision:!1,reasoning:!1,multilingual:!1,structured_output:!1},release_date:"2026-03-13",last_verified_at:"2026-06-22T12:36:35.292Z",release_date_source:"https://huggingface.co/BAAI/RoboBrain-Dex",release_date_verified_at:"2026-06-22T12:36:35.292Z",status:"draft"},{id:"baai/ursa-0.6b-ibq1024",name:"URSA-0.6B-IBQ1024",family:"BAAI",license:"apache-2.0",version:"0.6",provider:"BAAI",release_date:"2025-11-02",release_date_source:"https://huggingface.co/BAAI/URSA-0.6B-IBQ1024",huggingface_id:"BAAI/URSA-0.6B-IBQ1024",release_date_verified_at:"2026-06-23T11:42:38.141Z",last_verified_at:"2026-06-23T11:42:38.141Z",modality:"image-gen",specialized_architecture:{type:"diffusion",total_params_b:.864,notes:"Compact text-to-image diffusion model (URSA) using 

1an image byte-quantised (IBQ, 1024-codebook) tokenizer. 0.86B parameters, released open-weight by BAAI."},specialized_memory:{weights_bf16_gb:1.73,activation_estimate_gb:.6},specialized_pricing:{kind:"free_open_weights"},description:"URSA-0.6B-IBQ1024 is BAAI's compact open-weight text-to-image diffusion model built on an image byte-quantised (IBQ) tokenizer with a 1024-entry codebook. At 0.86B parameters it targets efficient image generation on modest hardware.",tags:["image-generation","diffusion","open-weight","text-to-image"]},{id:"baichuan/baichuan2-13b",name:"Baichuan 2 13B",provider:"Baichuan",family:"Baichuan 2",version:"2",release_date:"2023-09-25",architecture:{type:"dense",total_params_b:13,active_params_b:13,num_layers:40,hidden_dim:5120,num_heads:40,num_kv_heads:40,head_dim:128,intermediate_dim:13696,vocab_size:125696},context_window:4096,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:26,weights_fp8_gb:13,weights_int4_gb:6.5,kv_per_token_bytes:819200,activation_estimate_gb:1.5},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi"],api_pricing:{baichuan:{input_per_m:.25,output_per_m:.25}},tags:["chat","open-weight","mha"],huggingface_id:"baichuan-inc/Baichuan2-13B-Chat",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/baichuan-inc/Baichuan2-13B-Chat",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.178Z",license:"other"},{id:"baichuan/baichuan2-7b",name:"Baichuan 2 7B",provider:"Baichuan",family:"Baichuan 2",version:"2",release_date:"2023-09-06",architecture:{type:"dense",total_params_b:7,active_params_b:7,num_layers:32,hidden_dim:4096,num_heads:32,num_kv_heads:32,head_dim:128,intermediate_dim:11008,vocab_size:125696},context_window:4096,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:14,weights_fp8_gb:7,weights_int4_gb:3.5,kv_per_token_bytes:262144,activation_estimate_gb:.8},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","tgi"],api_pricing:{},tags:["chinese","multilingual","open-weight"],huggingface_id:"baichuan-inc/Baichuan2-7B-Chat",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/baichuan-inc/Baichuan2-7B-Chat",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-16T09:29:53.000Z"},{id:"bigcode/octocoder-15b",name:"OctoCoder 15B",provider:"BigCode",family:"OctoCoder",version:"1",release_date:"2023-08-25",architecture:{type:"dense",total_params_b:15.5,active_params_b:15.5,num_layers:40,hidden_dim:6144,num_heads:48,num_kv_heads:48,head_dim:128,intermediate_dim:24576,vocab_size:49152},context_window:8192,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:31,weights_fp8_gb:15.5,weights_int4_gb:7.75,kv_per_token_bytes:983040,activation_estimate_gb:1.2},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi"],api_pricing:{},tags:["code","code-specialized","open-weight"],huggingface_id:"bigcode/octocoder",huggingface_id_status:"manual",release_date_source:"https://huggingface.co/bigcode/octocoder",last_verified_at:"2026-06-25T16:19:04.532Z",license:"bigcode-openrail-m"},{id:"bigcode/santacoder-1.1b",name:"SantaCoder 1.1B",provider:"BigCode",family:"SantaCoder",version:"1",release_date:"2023-01-09",architecture:{type:"dense",total_params_b:1.1,active_params_b:1.1,num_layers:24,hidden_dim:2048,num_heads:16,num_kv_heads:16,head_dim:128,intermediate_dim:8192,vocab_size:49280},context_window:2048,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:2.2,weights_fp8_gb:1.1,weights_int4_gb:.55,kv_per_token_bytes:98304,activation_estimate_gb:.2},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","tgi"],api_pricing:{},tags:["code","code-specialized","open-weight","lightweight"],huggingface_id:"bigcode/santacoder",huggingface_id_status:"manual",release_date_source:"https://huggingface.co/bigcode/santacoder",last_verified_at:"2026-06-25T16:19:04.533Z",license:"bigcode-openrail-m"}

1,{id:"bigcode/starcoder2-15b",name:"StarCoder2 15B",provider:"BigCode",family:"StarCoder2",version:"2",release_date:"2024-02-28",architecture:{type:"dense",total_params_b:15.5,active_params_b:15.5,num_layers:40,hidden_dim:6144,num_heads:48,num_kv_heads:4,head_dim:128,intermediate_dim:24576,vocab_size:49152},context_window:16384,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:31,weights_fp8_gb:15.5,weights_int4_gb:7.75,kv_per_token_bytes:81920,activation_estimate_gb:1.5},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm"],api_pricing:{huggingface:{input_per_m:.3,output_per_m:.3}},tags:["code","open-weight","code-specialized"],huggingface_id:"bigcode/starcoder2-15b",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/bigcode/starcoder2-15b",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.178Z",license:"bigcode-openrail-m"},{id:"bigcode/starcoder2-3b",name:"StarCoder2 3B",provider:"BigCode",family:"StarCoder2",version:"2",release_date:"2024-02-28",architecture:{type:"dense",total_params_b:3.03,active_params_b:3.03,num_layers:30,hidden_dim:3072,num_heads:24,num_kv_heads:2,head_dim:128,intermediate_dim:12288,vocab_size:49152},context_window:16384,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:6.06,weights_fp8_gb:3.03,weights_int4_gb:1.515,kv_per_token_bytes:30720,activation_estimate_gb:.5},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","ollama"],api_pricing:{huggingface:{input_per_m:.1,output_per_m:.1}},tags:["code","open-weight","code-specialized"],huggingface_id:"bigcode/starcoder2-3b",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/bigcode/starcoder2-3b",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.178Z",license:"bigcode-openrail-m"},{id:"bigcode/starcoder2-7b",name:"StarCoder2 7B",provider:"BigCode",family:"StarCoder2",version:"2",release_date:"2024-02-28",architecture:{type:"dense",total_params_b:6.73,active_params_b:6.73,num_layers:32,hidden_dim:4608,num_heads:36,num_kv_heads:4,head_dim:128,intermediate_dim:18432,vocab_size:49152},context_window:16384,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:13.46,weights_fp8_gb:6.73,weights_int4_gb:3.365,kv_per_token_bytes:65536,activation_estimate_gb:1},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{huggingface:{input_per_m:.15,output_per_m:.15}},tags:["code","open-weight","code-specialized"],huggingface_id:"bigcode/starcoder2-7b",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/bigcode/starcoder2-7b",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.179Z",license:"bigcode-openrail-m"},{id:"black-forest-labs/flux-1-dev",name:"FLUX.1 Dev",provider:"Black Forest Labs",family:"FLUX",version:"1-dev",release_date:"2024-08-01",architecture:{type:"dense",total_params_b:12,active_params_b:12,num_layers:38,hidden_dim:3072,num_heads:24,num_kv_heads:24,head_dim:128,intermediate_dim:12288,vocab_size:49408},context_window:512,supported_precisions:["bf16","fp8"],default_precision:"bf16",memory:{weights_bf16_gb:24,weights_fp8_gb:12,weights_int4_gb:6,kv_per_token_bytes:175104,activation_estimate_gb:4},capabilities:{tools:!1,vision:!0,structured_output:!1,code:!1,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:[],api_pricing:{together:{input_per_m:25,output_per_m:25}},tags:["image-generation","open-weight"],huggingface_id:"Comfy-Org/flux1-dev",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/Comfy-Org/flux1-dev",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.179Z",license:"other"},{id:"black-forest-labs/flux-2",name:"FLUX.2",provider:"Black Forest Labs",fam

1ily:"FLUX",version:"2",release_date:"2025-11-25",release_date_source:"https://bfl.ai/blog/flux-2",release_date_verified_at:"2026-05-13T14:42:10Z",last_verified_at:"2026-05-13T14:42:10Z",architecture:{type:"dense",total_params_b:12,active_params_b:12,num_layers:28,hidden_dim:3072,num_heads:24,num_kv_heads:24,head_dim:128,intermediate_dim:8192,vocab_size:49408},context_window:4096,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:24,weights_fp8_gb:12,weights_int4_gb:6,kv_per_token_bytes:196608,activation_estimate_gb:1},capabilities:{tools:!1,vision:!0,structured_output:!1,code:!1,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","tensorrt-llm"],api_pricing:{},tags:["image-generation","open-weight"]},{id:"cerebras/btlm-3b",name:"BTLM 3B",provider:"Cerebras",family:"BTLM",version:"1",release_date:"2023-08-28",architecture:{type:"dense",total_params_b:3,active_params_b:3,num_layers:32,hidden_dim:2560,num_heads:32,num_kv_heads:32,head_dim:80,intermediate_dim:6912,vocab_size:50257},context_window:8192,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:6,weights_fp8_gb:3,weights_int4_gb:1.5,kv_per_token_bytes:163840,activation_estimate_gb:.3},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","tgi"],api_pricing:{},tags:["chat","code","open-weight","lightweight"],huggingface_id:"cerebras/btlm-3b-8k-base",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/cerebras/btlm-3b-8k-base",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.179Z",license:"apache-2.0"},{id:"cerebras/cerebras-gpt-13b",name:"Cerebras GPT 13B",provider:"Cerebras",family:"Cerebras GPT",version:"1",release_date:"2023-03-28",architecture:{type:"dense",total_params_b:13,active_params_b:13,num_layers:40,hidden_dim:5120,num_heads:40,num_kv_heads:40,head_dim:128,intermediate_dim:20480,vocab_size:50257},context_window:2048,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:26,weights_fp8_gb:13,weights_int4_gb:6.5,kv_per_token_bytes:409600,activation_estimate_gb:1},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","tgi"],api_pricing:{},tags:["base","open-weight"],huggingface_id:"cerebras/Cerebras-GPT-13B",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/cerebras/Cerebras-GPT-13B",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.179Z",license:"apache-2.0"},{id:"codellama/codellama-70b",name:"Code Llama 70B",provider:"Meta",family:"Code Llama",version:"1",release_date:"2024-01-29",architecture:{type:"dense",total_params_b:70,active_params_b:70,num_layers:80,hidden_dim:8192,num_heads:64,num_kv_heads:8,head_dim:128,intermediate_dim:28672,vocab_size:32016},context_window:16384,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:140,weights_fp8_gb:70,weights_int4_gb:35,kv_per_token_bytes:327680,activation_estimate_gb:2.5},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!0,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm"],api_pricing:{together:{input_per_m:.9,output_per_m:.9},fireworks:{input_per_m:.9,output_per_m:.9}},tags:["code","code-specialized","open-weight"],release_date_source:"https://www.llama.com/",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-16T09:29:53.000Z"},{id:"cognitivecomputations/dolphin-2.9-72b",name:"Dolphin 2.9 72B",provider:"Cognitive Computations",family:"Dolphin",version:"2.9",release_date:"2024-06-15",architecture:{type:"dense",total_params_b:72,active_params_b:72,num_layers:80,hidden_dim:8192,num_heads:64,num_kv_heads:8,head_dim:128,intermediate_dim:24576,vocab_size:152064},context_window:32768,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:144,weights_fp8_gb:72,weights_int4_gb:36,kv_per_token_bytes:327680,activation_estimate_gb:2.5},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm"],api_pricing:{},tags:["chat","code","instruction-following","open-weight","uncensored"],huggingface_id:"cognitivecomputations/dolphin-2.9-llama3-70b",huggingface_id_status:"manual",release_date_source:"https://huggingface.co/cognitivecomputations/dolphin-2.9-llama3-70b",last_verified_at:"2026-06-25T16:19:04.533Z",license:"llama3"},{id:"cohere/aya-23-35b",name:"Aya 23 35B",provider:"Cohere",family:"Aya",ver

1sion:"23",release_date:"2024-05-23",architecture:{type:"dense",total_params_b:35,active_params_b:35,num_layers:32,hidden_dim:8192,num_heads:64,num_kv_heads:8,head_dim:128,intermediate_dim:22528,vocab_size:256e3},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:70,weights_fp8_gb:35,weights_int4_gb:17.5,kv_per_token_bytes:131072,activation_estimate_gb:2},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm"],api_pricing:{cohere:{input_per_m:.5,output_per_m:1.5}},tags:["chat","multilingual","open-weight"],huggingface_id:"CohereLabs/aya-23-35B",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/CohereLabs/aya-23-35B",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.179Z",license:"cc-by-nc-4.0"},{id:"cohere/aya-23-8b",name:"Aya 23 8B",provider:"Cohere",family:"Aya",version:"23",release_date:"2024-05-23",architecture:{type:"dense",total_params_b:8,active_params_b:8,num_layers:32,hidden_dim:4096,num_heads:32,num_kv_heads:8,head_dim:128,intermediate_dim:14336,vocab_size:256e3},context_window:8192,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:16,weights_fp8_gb:8,weights_int4_gb:4,kv_per_token_bytes:131072,activation_estimate_gb:.8},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{cohere:{input_per_m:.2,output_per_m:.6}},tags:["chat","multilingual","open-weight"],huggingface_id:"CohereLabs/aya-23-8B",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/CohereLabs/aya-23-8B",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.179Z",license:"cc-by-nc-4.0"},{id:"cohere/command-r-08-2024",name:"Command R (August 2024)",provider:"Cohere",family:"Command R",version:"08-2024",release_date:"2024-08-15",architecture:{type:"dense",total_params_b:35,active_params_b:35,num_layers:40,hidden_dim:8192,num_heads:64,num_kv_heads:8,head_dim:128,intermediate_dim:22016,vocab_size:256e3},context_window:128e3,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:70,weights_fp8_gb:35,weights_int4_gb:17.5,kv_per_token_bytes:327680,activation_estimate_gb:2},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi"],api_pricing:{cohere:{input_per_m:.15,output_per_m:.6},openrouter:{input_per_m:.15,output_per_m:.6}},tags:["chat","code","rag","open-weight","estimated-architecture"],huggingface_id:"CohereLabs/c4ai-command-r-08-2024",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/CohereLabs/c4ai-command-r-08-2024",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.179Z",license:"cc-by-nc-4.0"},{id:"cohere/command-r-7b",name:"Command R 7B",provider:"Cohere",family:"Command R",version:"1",release_date:"2024-10-15",architecture:{type:"dense",total_params_b:7,active_params_b:7,num_layers:32,hidden_dim:4096,num_heads:32,num_kv_heads:8,head_dim:128,intermediate_dim:14336,vocab_size:256e3},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:14,weights_fp8_gb:7,weights_int4_gb:3.5,kv_per_token_bytes:131072,activation_estimate_gb:.8},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{cohere:{input_per_m:.075,output_per_m:.15}},tags:["chat","rag","multilingual","open-weight"],huggingface_id:"CohereLabs/c4ai-command-r7b-12-2024",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/CohereLabs/c4ai-command-r7b-12-2024",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.179Z",license:"cc-by-nc-4.0"},{id:"cohere/command-r-plus",name:"Command R+",provider:"Cohere",family:"Command R",version:"plus",release_date:"2024-04-04",architecture:{type:"dense",total_params_b:104,active_params_b:104,num_layers:64,hidden_dim:12288,num_heads:96,num_kv_heads:96,head_dim:128,intermediate_dim:33792,vocab_size:256e3},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:208,weights_fp8_gb:104,weights_int4_gb:52,kv_per_token_bytes:3145728,activation_estimate_gb:3},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm"],api_pricing:{cohere:{input_per_m:2.5,output_per_m:10},together:{input_per_m:2,output_per_m:2},openrouter:{input_per_m:2.5,output_per_m:10}},tags:["chat","rag","instruction-following","open-weight"],huggingface_id:"CohereLabs/c4ai-command-r-plus",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/CohereLabs/c4ai-command-r-plus",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.179Z",license:"cc-by-nc-4.0"},{id:"cohere/command-r",name:"Command R",provider:"Cohere",family:"Command R",version:"1",release_date:"2024-03-11",architecture:{type:"dense",total_params_b:35,active_params_b:35,num_layers:40,hidden_dim:8192,num_heads:64,num_kv_heads:8,head_dim:128,intermediate_dim:22528,vocab_size:256e3},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:70,weights_fp8_gb:35,weights_int4_gb:17.5,kv_per_token_bytes:163840,activation_estimate_gb:2},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm"],api_pricing:{cohere:{input_per_m:.5,output_per_m:1.5},together:{input_per_m:.5,output_per_m:.5}},tags:["chat","rag","instruction-following","open-weight"],huggingface_id:"CohereLabs/c4ai-command-r7b-12-2024",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/CohereLabs/c4ai-command-r7b-12-2024",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.179Z",license:"cc-by-nc-4.0"},{id:"cohere/embed-english-v3",name:"Cohere Embed English v3",provider:"Cohere",family:"Embed",version:"3",release_date:"2023-11-02",architecture:{type:"dense",total_params_b:.5,active_params_b:.5,num_layers:24,hidden_dim:1024,num_heads:16,num_kv_heads:16,head_dim:64,intermediate_dim:4096,vocab_size:256e3},context_window:512,supported_precisions:["bf16"],default_precision:"bf16",memory:{weights_bf16_gb:1,weights_fp8_gb:.5,weights_int4_gb:.25,kv_per_token_bytes:49152,activation_estimate_gb:.1},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:[],api_pricing:{cohere:{input_per_m:.1,output_per_m:.1}},tags:["embedding","proprietary","api-only","estimated-architecture"],huggingface_id:"CohereLabs/Cohere-embed-english-v3.0",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/CohereLabs/Cohere-embed-english-v3.0",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-16T09:29:53.000Z"},{id:"databricks/dbrx-base",name:"DBRX Base",provider:"Databricks",family:"DBRX",version:"1",release_date:"2024-03-27",architecture:{type:"moe",total_params_b:132,active_params_b:36,num_layers:40,hidden_dim:6144,num_heads:48,num_kv_heads:8,head_dim:128,intermediate_dim:10752,vocab_size:100352,num_experts:16,experts_per_token:4},context_window:32768,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:264,weights_fp8_gb:132,weights_int4_gb:66,kv_per_token_bytes:163840,activation_estimate_gb:2},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!0,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm"],api_pricing:{databricks:{input_per_m:.75,output_per_m:2.25}},tags:["base","moe","open-weight"],huggingface_id:"Undi95/dbrx-base",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/Undi95/dbrx-base",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.179Z",license:"other"}

1,{id:"dbrx/dbrx-instruct",name:"DBRX Instruct",provider:"Databricks",family:"DBRX",version:"1",release_date:"2024-03-27",architecture:{type:"moe",total_params_b:132,active_params_b:36,num_layers:40,hidden_dim:6144,num_heads:48,num_kv_heads:8,head_dim:128,intermediate_dim:10752,vocab_size:100352,num_experts:16,experts_per_token:4},context_window:32768,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:264,weights_fp8_gb:132,weights_int4_gb:66,kv_per_token_bytes:163840,activation_estimate_gb:2},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm"],api_pricing:{databricks:{input_per_m:.75,output_per_m:2.25},together:{input_per_m:1.2,output_per_m:1.2}},tags:["chat","code","moe","open-weight"],huggingface_id:"alpindale/dbrx-instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/alpindale/dbrx-instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-16T09:29:53.000Z"},{id:"deepseek/deepseek-coder-33b",name:"DeepSeek Coder 33B",provider:"DeepSeek",family:"DeepSeek Coder",version:"1",release_date:"2023-11-02",architecture:{type:"dense",total_params_b:33,active_params_b:33,num_layers:62,hidden_dim:7168,num_heads:56,num_kv_heads:8,head_dim:128,intermediate_dim:19200,vocab_size:32256},context_window:16384,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:66,weights_fp8_gb:33,weights_int4_gb:16.5,kv_per_token_bytes:253952,activation_estimate_gb:1.8},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!0,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm"],api_pricing:{together:{input_per_m:.8,output_per_m:.8}},tags:["code","code-specialized","open-weight"],huggingface_id:"deepseek-ai/deepseek-coder-33b-instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/deepseek-ai/deepseek-coder-33b-instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.179Z",license:"other"},{id:"deepseek/deepseek-coder-6.7b",name:"DeepSeek Coder 6.7B",provider:"DeepSeek",family:"DeepSeek Coder",version:"1",release_date:"2023-11-02",architecture:{type:"dense",total_params_b:6.7,active_params_b:6.7,num_layers:32,hidden_dim:4096,num_heads:32,num_kv_heads:32,head_dim:128,intermediate_dim:11008,vocab_size:32256},context_window:16384,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:13.4,weights_fp8_gb:6.7,weights_int4_gb:3.35,kv_per_token_bytes:524288,activation_estimate_gb:.8},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!0,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{together:{input_per_m:.2,output_per_m:.2}},tags:["code","code-specialized","open-weight"],huggingface_id:"deepseek-ai/deepseek-coder-6.7b-instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/deepseek-ai/deepseek-coder-6.7b-instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.179Z",license:"other"},{id:"deepseek/deepseek-coder-v2-236b",name:"DeepSeek Coder V2 236B",provider:"DeepSeek",family:"DeepSeek Coder V2",version:"2",release_date:"2024-06-17",architecture:{type:"moe",total_params_b:236,active_params_b:21,num_layers:60,hidden_dim:5120,num_heads:128,num_kv_heads:1,head_dim:128,intermediate_dim:12288,vocab_size:100015,num_experts:128,experts_per_token:6,shared_experts:2},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:472,weights_fp8_gb:236,weights_int4_gb:118,kv_per_token_bytes:30720,activation_estimate_gb:3},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tensorrt-llm"],api_pricing:{deepseek:{input_per_m:.14,output_per_m:.28},together:{input_per_m:.9,output_per_m:.9}},tags:["code","moe","code-specialized","open-weight"],release_date_source:"https://api-docs.deepseek.com/quick_start/pricing",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-16T09:29:53.000Z"},{id:"deepseek/deepseek-llm-67b",name:"DeepSeek LLM 67B",provider:"DeepSeek",family:"DeepSeek LLM",version:"1",release_date:"2024-01-05",architecture:{type:"dense",total_params_b:67,active_params_b:67,num_layers:95,hidden_dim:8192,num_heads:64,num_kv_heads:8,head_dim:128,intermediate_dim:22016,vocab_size:102400},context_window:4096,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:134,weights_fp8_gb:67,weights_int4_gb:33.5,kv_per_token_bytes:655360,activation_estimate_gb:2.5},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm"],api_pricing:{},tags:["chat","code","open-weight"],huggingface_id:"deepseek-ai/deepseek-llm-67b-base",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/deepseek-ai/deepseek-llm-67b-base",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.179Z",license:"other"},{id:"deepseek/deepseek-math-7b",name:"DeepSeek Math 7B",provider:"DeepSeek",family:"DeepSeek Math",version:"1",release_date:"2024-02-05",architecture:{type:"dense",total_params_b:7.24,active_params_b:7.24,num_layers:30,hidden_dim:4096,num_heads:32,num_kv_heads:32,head_dim:128,intermediate_dim:11008,vocab_size:102400},context_window:4096,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:14.48,weights_fp8_gb:7.24,weights_int4_gb:3.62,kv_per_token_bytes:491520,activation_estimate_gb:.8},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!0,multilingual:!1,reasoning:!0},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{},tags:["math","math-specialized","reasoning","open-weight"],huggingface_id:"deepseek-ai/deepseek-math-7b-instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/deepseek-ai/deepseek-math-7b-instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.179Z",license:"other"},{id:"deepseek/deepseek-moe-16b",name:"DeepSeek MoE 16B",provider:"DeepSeek",family:"DeepSeek MoE",version:"1",release_date:"2024-01-11",architecture:{type:"moe",total_params_b:16.4,active_params_b:2.8,num_layers:28,hidden_dim:2048,num_heads:16,num_kv_heads:16,head_dim:128,intermediate_dim:5632,vocab_size:102400,num_experts:64,experts_per_token:6},context_window:4096,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:32.8,weights_fp8_gb:16.4,weights_int4_gb:8.2,kv_per_token_bytes:229376,activation_estimate_gb:.5},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!0,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi"],api_pricing:{},tags:["chat","code","moe","open-weight"],huggingface_id:"deepseek-ai/deepseek-moe-16b-base",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/deepseek-ai/deepseek-moe-16b-base",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.179Z",license:"other"},{id:"deepseek/deepseek-r1-distill-1.5b",name:"DeepSeek R1 Distill 1.5B",provider:"DeepSeek",family:"DeepSeek R1",version:"1-distill",release_date:"2025-01-20",architecture:{type:"dense",total_params_b:1.5,active_params_b:1.5,num_layers:28,hidden_dim:1536,num_heads:12,num_kv_heads:2,head_dim:128,intermediate_dim:8960,vocab_size:151936},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:3,weights_fp8_gb:1.5,weights_int4_gb:.75,kv_per_token_bytes:14336,activation_estimate_gb:.2},capabilities:{tools:!1,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm","sglang","tgi","ollama"],api_pricing:{},tags:["chat","code","reasoning","distilled","open-weight","lightweight"],huggingface_id:"deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.179Z",license:"mit"},{id:"deepseek/deepseek-r1-distill-14b",name:"DeepSeek R1 Distill 14B",provider:"DeepSeek",family:"DeepSeek R1",version:"1-distill",release_date:"2025-01-20",architecture:{type:"dense",total_params_b:14.8,active_params_b:14.8,num_layers:48,hidden_dim:5120,num_heads:40,num_kv_heads:8,head_dim:128,intermediate_dim:13824,vocab_size:152064},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:29.6,weights_fp8_gb:14.8,weights_int4_gb:7.4,kv_per_token_bytes:196608,activation_estimate_gb:1.5},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{together:{input_per_m:.3,output_per_m:.3}},tags:["chat","code","reasoning","distilled","open-weight"],huggingface_id:"deepseek-ai/DeepSeek-R1-Distill-Qwen-14B",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-14B",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.179Z",license:"mit"},{id:"deepseek/deepseek-r1-distill-32b",name:"DeepSeek R1 Distill 32B",provider:"DeepSeek",family:"DeepSeek R1",version:"1-distill",release_date:"2025-01-20",architecture:{type:"dense",total_params_b:32.8,active_params_b:32.8,num_layers:64,hidden_dim:5120,num_heads:40,num_kv_heads:8,head_dim:128,intermediate_dim:25600,vocab_size:152064},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:65.6,weights_fp8_gb:32.8,weights_int4_gb:16.4,kv_per_token_bytes:262144,activation_estimate_gb:2},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{together:{input_per_m:.6,output_per_m:.6},fireworks:{input_per_m:.5,output_per_m:.5},openrouter:{input_per_m:.29,output_per_m:.29}},tags:["chat","code","reasoning","distilled","open-weight"],huggingface_id:"deepseek-ai/DeepSeek-R1-Distill-Qwen-32B",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-32B",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.179Z",license:"mit"},{id:"deepseek/deepseek-r1-distill-70b",name:"DeepSeek R1 Distill 70B",provider:"DeepSeek",family:"DeepSeek R1",version:"1-distill",release_date:"2025-01-20",architecture:{type:"dense",total_params_b:70.6,active_params_b:70.6,num_layers:80,hidden_dim:8192,num_heads:64,num_kv_heads:8,head_dim:128,intermediate_dim:28672,vocab_size:128256},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:14

11.2,weights_fp8_gb:70.6,weights_int4_gb:35.3,kv_per_token_bytes:327680,activation_estimate_gb:2.5},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm"],api_pricing:{together:{input_per_m:.88,output_per_m:.88},fireworks:{input_per_m:.9,output_per_m:.9},novita:{input_per_m:.8,output_per_m:.8},scaleway:{input_per_m:0,output_per_m:0},featherless:{input_per_m:0,output_per_m:0},openrouter:{input_per_m:.7,output_per_m:.7999999999999999}},tags:["chat","code","reasoning","distilled","open-weight"],huggingface_id:"deepseek-ai/DeepSeek-R1-Distill-Llama-70B",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Llama-70B",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.179Z",license:"mit"},{id:"deepseek/deepseek-r1-distill-8b",name:"DeepSeek R1 Distill 8B",provider:"DeepSeek",family:"DeepSeek R1",version:"1-distill",release_date:"2025-01-20",architecture:{type:"dense",total_params_b:8,active_params_b:8,num_layers:32,hidden_dim:4096,num_heads:32,num_kv_heads:8,head_dim:128,intermediate_dim:14336,vocab_size:128256},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:16,weights_fp8_gb:8,weights_int4_gb:4,kv_per_token_bytes:131072,activation_estimate_gb:1},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{together:{input_per_m:.2,output_per_m:.2}},tags:["chat","code","reasoning","distilled","open-weight"],huggingface_id:"deepseek-ai/DeepSeek-R1-0528-Qwen3-8B",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/deepseek-ai/DeepSeek-R1-0528-Qwen3-8B",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.180Z",license:"mit"},{id:"deepseek/deepseek-r1",name:"DeepSeek R1",provider:"DeepSeek",family:"DeepSeek R1",version:"1",release_date:"2025-01-20",architecture:{type:"moe",total_params_b:671,active_params_b:37,num_layers:61,hidden_dim:7168,num_heads:128,num_kv_heads:1,head_dim:128,intermediate_dim:18432,vocab_size:129280,num_experts:256,experts_per_token:8,shared_experts:1},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:1342,weights_fp8_gb:671,weights_int4_gb:335.5,kv_per_token_bytes:31232,activation_estimate_gb:3},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm","sglang","tensorrt-llm"],api_pricing:{deepseek:{input_per_m:.55,output_per_m:2.19},together:{input_per_m:3,output_per_m:7},novita:{input_per_m:.7,output_per_m:2.5},hyperbolic:{input_per_m:2,output_per_m:2},openrouter:{input_per_m:.7,output_per_m:2.5}},tags:["chat","code","reasoning","moe","open-weight","flagship"],huggingface_id:"deepseek-ai/DeepSeek-R1",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/deepseek-ai/DeepSeek-R1",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.180Z",license:"mit"},{id:"deepseek/deepseek-v2-lite",name:"DeepSeek V2 Lite",provider:"DeepSeek",family:"DeepSeek V2",version:"2-lite",release_date:"2024-05-16",architecture:{type:"moe",total_params_b:15.7,active_params_b:2.4,num_layers:27,hidden_dim:2048,num_heads:16,num_kv_heads:16,head_dim:128,intermediate_dim:10944,vocab_size:102400,num_experts:64,experts_per_token:6},context_window:32768,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:31.4,weights_fp8_gb:15.7,weights_int4_gb:7.85,kv_per_token_bytes:221184,activation_estimate_gb:.5},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi"],api_pricing:{},tags:["chat","code","moe","open-weight"],huggingface_id:"deepseek-ai/DeepSeek-V2-Lite-Chat",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/deepseek-ai/DeepSeek-V2-Lite-Chat",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.180Z",license:"other"},{id:"deepseek/deepseek-v2.5",name:"DeepSeek V2.5",provider:"DeepSeek",family:"DeepSeek V2",version:"2.5",release_date:"2024-09-05",architecture:{type:"moe",total_params_b:236,active_params_b:21,num_layers:60,hidden_dim:5120,num_heads:128,num_kv_heads:1,head_dim:128,intermediate_dim:12288,vocab_size:100015,num_experts:128,experts_per_token:6,shared_experts:2},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:472,weights_fp8_gb:236,weights_int4_gb:118,kv_per_token_bytes:30720,activation_estimate_gb:3},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tensorrt-llm"],api_pricing:{deepseek:{input_per_m:.14,output_per_m:.28},together:{input_per_m:.8,output_per_m:.8}},tags:["chat","code","moe","open-weight"],huggingface_id:"deepseek-ai/DeepSeek-V2.5",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/deepseek-ai/DeepSeek-V2.5",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.180Z",license:"other"},{id:"deepseek/deepseek-v3-0324",name:"DeepSeek V3-0324",provider:"DeepSeek",family:"DeepSeek V3",version:"3-0324",release_date:"2025-03-24",architecture:{type:"moe",total_params_b:685,active_params_b:37,num_layers:61,hidden_dim:7168,num_heads:128,num_kv_heads:1,head_dim:128,intermediate_dim:18432,vocab_size:129280,num_experts:256,experts_per_token:8,shared_experts:1},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"fp8",memory:{weights_bf16_gb:1370,weights_fp8_gb:685,weights_int4_gb:342.5,kv_per_token_bytes:31232,activation_estimate_gb:3},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm","sglang","tensorrt-llm"],api_pricing:{deepseek:{input_per_m:.28,output_per_m:.42},together:{input_per_m:1,output_per_m:1},novita:{input_per_m:.27,output_per_m:1.12},hyperbolic:{input_per_m:1.25,output_per_m:1.25},featherless:{input_per_m:0,output_per_m:0},openrouter:{input_per_m:.2,output_per_m:.77},fireworks:{input_per_m:.9,output_per_m:.9},deepinfra:{input_per_m:.4,output_per_m:.85}},tags:["chat","code","reasoning","moe","open-weight","flagship"],huggingface_id:"deepseek-ai/DeepSeek-V3-0324",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/deepseek-ai/DeepSeek-V3-0324",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.180Z",license:"mit"},{id:"deepseek/deepseek-v3",name:"DeepSeek V3",provider:"DeepSeek",family:"DeepSeek V3",version:"3",release_date:"2024-12-26",architecture:{type:"moe",total_params_b:671,active_params_b:37,num_layers:61,hidden_dim:7168,num_heads:128,num_kv_heads:1,head_dim:128,intermediate_dim:18432,vocab_size:129280,num_experts:256,experts_per_token:8,shared_experts:1},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:1342,weights_fp8_gb:671,weights_int4_gb:335.5,kv_per_token_bytes:31232,activation_estimate_gb:3},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tensorrt-llm"],api_pricing:{deepseek:{input_per_m:.28,output_per_m:.42},together:{input_per_m:.5,output_per_m:2.8},novita:{input_per_m:.4,output_per_m:1.3},openrouter:{input_per_m:.32,output_per_m:.8899999999999999},"together-ai":{input_per_m:.5,output_per_m:2.8}},tags:["chat","code","moe","open-weight","flagship"],huggingface_id:"deepseek-ai/DeepSeek-V3",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/deepseek-ai/DeepSeek-V3",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-18T13:16:54.244Z"},{id:"deepseek/deepseek-v4-flash",name:"DeepSeek V4 Flash",family:"DeepSeek",provider:"DeepSeek",is_partial:!0,release_date:"2026-04-24",release_date_source:"https://openrouter.ai/models/deepseek/deepseek-v4-flash",release_date_verified_at:"2026-06-20T16:00:00Z",last_verified_at:"2026-06-20T16:00:00Z",context_window:1048576,deployment:"api",api_pricing:{deepseek:{input_per_m:.14,output_per_m:.28}},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},description:"DeepSeek V4 Flash is the low-cost, high-speed tier of DeepSeek's V4 API family — reasoning and tool use with a ~1M-token context window at some of the cheapest per-token pricing available."},{id:"deepseek/deepseek-v4-pro",name:"DeepSeek V4 Pro",family:"DeepSeek",provider:"DeepSeek",is_partial:!0,release_date:"2026-04-24",release_date_source:"https://openrouter.ai/models/deepseek/deepseek-v4-pro",release_date_verified_at:"2026-06-20T16:00:00Z",last_verified_at:"2026-06-20T16:00:00Z",context_window:1048576,deployment:"api",api_pricing:{deepseek:{input_per_m:.435,output_per_m:.87}},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},description:"DeepSeek V4 Pro is the high-capability tier of DeepSeek's V4 API family — a reasoning-focused, tool-using model with a ~1M-token context window at aggressively low per-token pricing."},{id:"deepseek/janus-pro-7b",name:"Janus-Pro 7B",provider:"DeepSeek",family:"Janus",version:"Pro",release_date:"2025-01-27",architecture:{type:"dense",total_params_b:7,active_params_b:7,num_layers:32,hidden_dim:4096,num_heads:32,num_kv_heads:32,head_dim:128,intermediate_dim:11008,vocab_size:1e5},context_window:8192,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:14,weights_fp8_gb:7,weights_int4_gb:3.5,kv_per_token_bytes:262144,activation_estimate_gb:.8},capabilities:{tools:!1,vision:!0,structured_output:!1,code:!0,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","ollama"],api_pricing:{},tags:["vision","image-generation","open-weight"],huggingface_id:"deepseek-ai/Janus-Pro-7B",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/deepseek-ai/Janus-Pro-7B",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.180Z",license:"mit"},{id:"falcon/falcon-180b",name:"Falcon 180B",provider:"TII",family:"Falcon",version:"1",release_date:"2023-09-06",architecture:{type:"dense",total_params_b:180,active_params_b:180,num_layers:80,hidden_dim:14848,num_heads:232,num_kv_heads:8,head_dim:64,intermediate_dim:59392,vocab_size:65024},context_window:2048,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:360,weights_fp8_gb:180,weights_int4_gb:90,kv_per_token_bytes:163840,activation_estimate_gb:4},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm"],api_pricing:{tii:{input_per_m:2.4,output_per_m:2.4}},tags:["chat","open-weight"],huggingface_id:"tiiuae/falcon-180B-chat",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/tiiuae/falcon-180B-chat",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.180Z",license:"unknown"},{id:"falcon/falcon-40b",name:"Falcon 40B",provider:"TII",family:"Falcon",version:"1",release_date:"2023-05-25",architecture:{type:"dense",total_params_b:40,active_params_b:40,num_layers:60,hidden_dim:8192,num_heads:128,num_kv_heads:8,head_dim:64,intermediate_dim:32768,vocab_size:65024},context_window:2048,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:80,weights_fp8_gb:40,weights_int4_gb:20,kv_per_token_bytes:122880,activation_estimate_gb:2},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm"],api_pricing:{tii:{input_per_m:.8,output_per_m:.8}},tags:["chat","open-weight"],huggingface_id:"tiiuae/falcon-40b",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/tiiuae/falcon-40b",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.180Z",license:"apache-2.0"},{id:"falcon/falcon-7b",name:"Falcon 7B",provider:"TII",family:"Falcon",version:"1",release_date:"2023-05-25",architecture:{type:"dense",total_params_b:7,active_params_b:7,num_layers:32,hidden_dim:4544,num_heads:71,num_kv_heads:1,head_dim:64,intermediate_dim:18176,vocab_size:65024},context_window:2048,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:14,weights_fp8_gb:7,weights_int4_gb:3.5,kv_per_token_bytes:8192,activation_estimate_gb:.5},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{tii:{input_per_m:.15,output_per_m:.15}},tags:["chat","open-weight","mqa"],huggingface_id:"tiiuae/falcon-7b",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/tiiuae/falcon-7b",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.180Z",license:"apache-2.0"},{id:"finance/fingpt-7b",name:"FinGPT 7B",provider:"AI4Finance",family:"FinGPT",version:"1",release_date:"2023-10-15",architecture:{type:"dense",total_params_b:7.2,active_params_b:7.2,num_layers:32,hidden_dim:4096,num_heads:32,num_kv_heads:32,head_dim:128,intermediate_dim:11008,vocab_size:32e3},context_window:4096,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:14.4,weights_fp8_gb:7.2,weights_int4_gb:3.6,kv_per_token_bytes:262144,activation_estimate_gb:.8},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!0,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","tgi","ollama"],api_pricing:{},tags:["finance","domain-specific","open-weight"],huggingface_id:"dohonba/mistral_7b_fingpt",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/dohonba/mistral_7b_fingpt",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.180Z",license:"apache-2.0"},{id:"google/codegemma-7b",name:"CodeGemma 7B",provider:"Google",family:"Gemma",version:"1",release_date:"2024-04-09",architecture:{type:"dense",total_params_b:8.5,active_params_b:8.5,num_layers:28,hidden_dim:3072,num_heads:16,num_kv_heads:16,head_dim:256,intermediate_dim:24576,vocab_size:256128},context_window:8192,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:17,weights_fp8_gb:8.5,weights_int4_gb:4.25,kv_per_token_bytes:229376,activation_estimate_gb:.8},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!0,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","ollama"],api_pricing:{},tags:["code","code-specialized","open-weight"],huggingface_id:"google/codegemma-7b-it",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/google/codegemma-7b-it",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.180Z",license:"gemma"},{id:"google/gemini-3.1-flash-image",name:"Nano Banana 2 (Gemini 3.1 Flash Image)",family:"Gemini",provider:"Google",release_date:"2026-06-18",release_date_source:"https://openrouter.ai/models/google/gemini-3.1-flash-image",release_date_verified_at:"2026-06-20T16:00:00Z",last_verified_at:"2026-06-20T16:00:00Z",modality:"image-gen",specialized_architecture:{type:"diffusion",notes:"Closed Google image-generation model (Nano Banana 2 / Gemini 3.1 Flash Image). Fast, low-cost image generation and editing through the Gemini API. Architecture and parameter count are not disclosed by Google."},specialized_pricing:{kind:"per_image",price:.067,resolution:"1K (0.5K: $0.045, 4K: $0.151)"},description:"Gemini 3.1 Flash Image (Nano Banana 2) is Google's fast, low-cost closed image-generation model, served through the Gemini API. Standard pricing is about $0.067 per 1K image.",tags:["image-generation","closed","gemini","nano-banana

1","google"]},{id:"google/gemini-3.1-flash-lite",name:"Gemini 3.1 Flash Lite",family:"Gemini",provider:"Google",is_partial:!0,release_date:"2026-05-07",release_date_source:"https://openrouter.ai/models/google/gemini-3.1-flash-lite",release_date_verified_at:"2026-06-20T16:00:00Z",last_verified_at:"2026-06-20T16:00:00Z",context_window:1048576,deployment:"api",api_pricing:{google:{input_per_m:.25,output_per_m:1.5}},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},description:"Gemini 3.1 Flash-Lite is Google's fastest, lowest-cost Gemini tier — multimodal and tool-capable with a ~1M-token context window, priced for high-throughput use through the Gemini API."},{id:"google/gemini-3.5-flash",name:"Gemini 3.5 Flash",family:"Gemini",provider:"Google",is_partial:!0,release_date:"2026-05-19",release_date_source:"https://openrouter.ai/models/google/gemini-3.5-flash",release_date_verified_at:"2026-06-20T16:00:00Z",last_verified_at:"2026-06-20T16:00:00Z",context_window:1048576,deployment:"api",api_pricing:{google:{input_per_m:1.5,output_per_m:9,reasoning_per_m:9}},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},description:"Google's Gemini 3.5 Flash is a fast, multimodal model with a ~1M-token context window and a dedicated thinking (reasoning) mode, priced for high-throughput use through the Gemini API."},{id:"google/gemini-3-pro-image",name:"Nano Banana Pro (Gemini 3 Pro Image)",family:"Gemini",provider:"Google",release_date:"2026-06-18",release_date_source:"https://openrouter.ai/models/google/gemini-3-pro-image",release_date_verified_at:"2026-06-20T16:00:00Z",last_verified_at:"2026-06-20T16:00:00Z",modality:"image-gen",specialized_architecture:{type:"diffusion",notes:"Closed Google image-generation model (Nano Banana Pro / Gemini 3 Pro Image). Generates and edits high-resolution images through the Gemini API. Architecture and parameter count are not disclosed by Google."},specialized_pricing:{kind:"per_image",price:.134,resolution:"1K/2K (4K: $0.24)"},description:"Gemini 3 Pro Image (Nano Banana Pro) is Google's flagship closed image-generation and editing model, served through the Gemini API. Standard pricing is $0.134 per 1K/2K image ($0.24 per 4K image).",tags:["image-generation","closed","gemini","nano-banana","google"]},{id:"google/gemini-3",name:"Gemini 3 Pro Preview",provider:"Google DeepMind",family:"Gemini",version:"3",release_date:"2026-04-09",release_date_source:"https://ai.google.dev/gemini-api/docs/pricing",release_date_verified_at:"2026-05-16T14:43:00Z",last_verified_at:"2026-05-16T14:43:00Z",is_partial:!1,tags:["frontier","multimodal"],context_window:1e6,supported_precisions:["bf16"],default_precision:"bf16",architecture:{type:"moe",total_params_b:600,active_params_b:100,num_layers:80,hidden_dim:10240,num_heads:80,num_kv_heads:10,head_dim:128,intermediate_dim:36864,vocab_size:2e5,num_experts:16,experts_per_token:2},memory:{weights_bf16_gb:1200,weights_fp8_gb:600,weights_int4_gb:300,kv_per_token_bytes:204800,activation_estimate_gb:4},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm"],api_pricing:{google:{input_per_m:2,output_per_m:12}}},{id:"google/gemma-1.1-2b",name:"Gemma 1.1 2B",provider:"Google",family:"Gemma",version:"1.1",release_date:"2024-04-09",architecture:{type:"dense",total_params_b:2.5,active_params_b:2.5,num_layers:18,hidden_dim:2048,num_heads:8,num_kv_heads:1,head_dim:256,intermediate_dim:16384,vocab_size:256e3},context_window:8192,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:5,weights_fp8_gb:2.5,weights_int4_gb:1.25,kv_per_token_bytes:18432,activation_estimate_gb:.2},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{},tags:["chat","lightweight","edge","open-weight"],huggingface_id:"google/gemma-1.1-2b-it",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/google/gemma-1.1-2b-it",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.180Z",license:"gemma"},{id:"google/gemma-2-27b",name:"Gemma 2 27B",provider:"Google",family:"Gemma 2",version:"2",release_date:"2024-06-27",architecture:{type:"dense",total_params_b:27,active_params_b:27,num_layers:46,hidden_dim:4608,num_heads:32,num_kv_heads:16,head_dim:128,intermediate_dim:36864,vocab_size:256e3},context_window:8192,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:54,weights_fp8_gb:27,weights_int4_gb:13.5,kv_per_token_bytes:376832,activation_estimate_gb:1.5},capabilities:{tools:!1,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{together:{input_per_m:.3,output_per_m:.3},deepinfra:{input_per_m:.27,output_per_m:.27},featherless:{input_per_m:0,output_per_m:0},scaleway:{input_per_m:0,output_per_m:0},openrouter:{input_per_m:.65,output_per_m:.65}},tags:["chat","code","open-weight"],huggingface_id:"google/gemma-3-27b-it",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/google/gemma-3-27b-it",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.180Z",license:"gemma"},{id:"google/gemma-2-2b",name:"Gemma 2 2B",provider:"Google",family:"Gemma 2",version:"2",release_date:"2024-0

17-31",architecture:{type:"dense",total_params_b:2.6,active_params_b:2.6,num_layers:26,hidden_dim:2304,num_heads:8,num_kv_heads:4,head_dim:256,intermediate_dim:9216,vocab_size:256e3},context_window:8192,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:5.2,weights_fp8_gb:2.6,weights_int4_gb:1.3,kv_per_token_bytes:106496,activation_estimate_gb:.3},capabilities:{tools:!1,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{},tags:["chat","code","open-weight","lightweight"],huggingface_id:"google/gemma-4-E2B-it",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/google/gemma-4-E2B-it",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.180Z",license:"apache-2.0"},{id:"google/gemma-2-9b",name:"Gemma 2 9B",provider:"Google",family:"Gemma 2",version:"2",release_date:"2024-06-27",architecture:{type:"dense",total_params_b:9.2,active_params_b:9.2,num_layers:42,hidden_dim:3584,num_heads:16,num_kv_heads:8,head_dim:256,intermediate_dim:14336,vocab_size:256e3},context_window:8192,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:18.4,weights_fp8_gb:9.2,weights_int4_gb:4.6,kv_per_token_bytes:344064,activation_estimate_gb:1},capabilities:{tools:!1,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{together:{input_per_m:.2,output_per_m:.2},deepinfra:{input_per_m:.1,output_per_m:.1},groq:{input_per_m:.2,output_per_m:.2}},tags:["chat","code","open-weight"],huggingface_id:"google/gemma-2-9b-it",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/google/gemma-2-9b-it",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.180Z",license:"gemma"},{id:"google/gemma-3-12b",name:"Gemma 3 12B",provider:"Google",family:"Gemma 3",version:"3",release_date:"2025-03-12",architecture:{type:"dense",total_params_b:12,active_params_b:12,num_layers:48,hidden_dim:3072,num_heads:32,num_kv_heads:16,head_dim:128,intermediate_dim:12288,vocab_size:262144},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:24,weights_fp8_gb:12,weights_int4_gb:6,kv_per_token_bytes:393216,activation_estimate_gb:1},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{google:{input_per_m:.05,output_per_m:.1},together:{input_per_m:.15,output_per_m:.15},openrouter:{input_per_m:.04,output_per_m:.13}},tags:["chat","code","multimodal","open-weight"],huggingface_id:"google/gemma-3-12b-it",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/google/gemma-3-12b-it",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.180Z",license:"gemma"},{id:"google/gemma-3-1b",name:"Gemma 3 1B",provider:"Google",family:"Gemma 3",version:"3",release_date:"2025-03-12",architecture:{type:"dense",total_params_b:1,active_params_b:1,num_layers:26,hidden_dim:1536,num_heads:16,num_kv_heads:4,head_dim:128,intermediate_dim:6144,vocab_size:262144},context_window:32768,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:2,weights_fp8_gb:1,weights_int4_gb:.5,kv_per_token_bytes:26624,activation_estimate_gb:.2},capabilities:{tools:!1,vision:!1,structured_output:!0,code:!0,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{},tags:["chat","open-weight","lightweight"],huggingface_id:"google/gemma-3-1b-it",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/google/gemma-3-1b-it",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.180Z",license:"gemma"},{id:"google/gemma-3-27b",name:"Gemma 3 27B",provider:"Google",family:"Gemma 3",version:"3",release_date:"2025-03-12",architecture:{type:"dense",total_params_b:27,active_params_b:27,num_layers:62,hidden_dim:3584,num_heads:32,num_kv_heads:16,head_dim:128,intermediate_dim:14336,vocab_size:262144},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:54,weights_fp8_gb:27,weights_int4_gb:13.5,kv_per_token_bytes:507904,activation_estimate_gb:1.5},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{google:{input_per_m:.1,output_per_m:.2},together:{input_per_m:.3,output_per_m:.3},featherless:{input_per_m:0,output_per_m:0},scaleway:{input_per_m:0,output_per_m:0},openrouter:{input_per_m:.08,output_per_m:.16}},tags:["chat","code","multimodal","open-weight"],huggingface_id:"google/gemma-3-27b-it",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/google/gemma-3-27b-it",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.181Z",license:"gemma"},{id:"google/gemma-3-2b",name:"Gemma 3 2B",provider:"Google",family:"Gemma 3",version:"3",release_date:"2025-03-12",architecture:{type:"dense",total_params_b:2,active_params_b:2,num_layers:26,hidden_dim:2304,num_heads:8,num_kv_heads:4,head_dim:256,intermediate_dim:9216,vocab_size:262144},context_window:8192,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:4,weights_fp8_gb:2,weights_int4_gb:1,kv_per_token_bytes:26624,activation_estimate_gb:.3},capabilities:{tools:!1,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","ollama"],api_pricing:{},tags:["chat","code","open-weight","lightweight"],huggingface_id:"google/gemma-3n-E2B-it",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/google/gemma-3n-E2B-it",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.181Z",license:"gemma"},{id:"google/gemma-3-4b",name:"Gemma 3 4B",provider:"Google",family:"Gemma 3",version:"3",release_date:"2025-03-12",architecture:{type:"dense",total_params_b:4.3,active_params_b:4.3,num_layers:34,hidden_dim:2560,num_heads:32,num_kv_heads:8,head_dim:128,intermediate_dim:10240,vocab_size:262144},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:8.6,weights_fp8_gb:4.3,weights_int4_gb:2.15,kv_per_token_bytes:139264,activation_estimate_gb:.5},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{google:{input_per_m:.05,output_per_m:.1},openrouter:{input_per_m:.04,output_per_m:.08}},tags:["chat","code","multimodal","open-weight"],huggingface_id:"google/gemma-3-4b-it",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/google/gemma-3-4b-it",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.181Z",license:"gemma"},{id:"google/gemma-4-31b",name:"Gemma 4 31B-IT",provider:"Google",family:"Gemma 4",version:"4",release_date:"2025-06-25",architecture:{type:"dense",total_params_b:31,active_params_b:31,num_layers:62,hidden_dim:4096,num_heads:32,num_kv_heads:16,head_dim:128,intermediate_dim:16384,vocab_size:262144},context_window:32768,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:62,weights_fp8_gb:31,weights_int4_gb:15.5,kv_per_token_bytes:507904,activation_estimate_gb:1.8},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{google:{input_per_m:.15,output_per_m:.3},together:{input_per_m:.2,output_per_m:.5},novita:{input_per_m:.14,output_per_m:.4},deepinfra:{input_per_m:.13,output_per_m:.38},featherless:{input_per_m:0,output_per_m:0},openrouter:{input_per_m:.12,output_per_m:.37}},tags:["chat","code","reasoning","open-weight"],huggingface_id:"google/gemma-4-31B-it",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/google/gemma-4-31B-it",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.181Z",license:"apache-2.0"}

1,{id:"google/paligemma-3b",name:"PaLI-Gemma 3B",provider:"Google",family:"PaLI-Gemma",version:"1",release_date:"2024-05-14",architecture:{type:"dense",total_params_b:2.9,active_params_b:2.9,num_layers:18,hidden_dim:2048,num_heads:8,num_kv_heads:8,head_dim:256,intermediate_dim:16384,vocab_size:257152},context_window:8192,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:5.8,weights_fp8_gb:2.9,weights_int4_gb:1.45,kv_per_token_bytes:73728,activation_estimate_gb:.4},capabilities:{tools:!1,vision:!0,structured_output:!1,code:!1,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","tgi"],api_pricing:{},tags:["vision","multimodal","open-weight"],huggingface_id:"google/paligemma-3b-pt-224",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/google/paligemma-3b-pt-224",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.181Z",license:"gemma"},{id:"google/recurrentgemma-2b",name:"RecurrentGemma 2B",provider:"Google",family:"RecurrentGemma",version:"1",release_date:"2024-04-09",architecture:{type:"dense",total_params_b:2.7,active_params_b:2.7,num_layers:26,hidden_dim:2560,num_heads:10,num_kv_heads:1,head_dim:256,intermediate_dim:7680,vocab_size:256e3},context_window:8192,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:5.4,weights_fp8_gb:2.7,weights_int4_gb:1.35,kv_per_token_bytes:26624,activation_estimate_gb:.3},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang"],api_pricing:{},tags:["chat","open-weight","lightweight","hybrid-architecture","estimated-architecture"],huggingface_id:"google/recurrentgemma-2b",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/google/recurrentgemma-2b",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.181Z",license:"gemma"},{id:"gradient/llama-3-70b-1m",name:"Llama 3 70B 1M Context",provider:"Gradient",family:"Llama 3",version:"3-1m",release_date:"2024-06-10",architecture:{type:"dense",total_params_b:70.6,active_params_b:70.6,num_layers:80,hidden_dim:8192,num_heads:64,num_kv_heads:8,head_dim:128,intermediate_dim:28672,vocab_size:128256},context_window:1048576,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:141.2,weights_fp8_gb:70.6,weights_int4_gb:35.3,kv_per_token_bytes:327680,activation_estimate_gb:2.5},capabilities:{tools:!1,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang"],api_pricing:{gradient:{input_per_m:1.5,output_per_m:1.5},openrouter:{input_per_m:.51,output_per_m:.74}},tags:["chat","code","long-context","open-weight"],huggingface_id:"gradientai/Llama-3-70B-Instruct-Gradient-1048k",huggingface_id_status:"manual",release_date_source:"https://huggingface.co/gradientai/Llama-3-70B-Instruct-Gradient-1048k",last_verified_at:"2026-06-25T16:19:04.533Z",license:"llama3"},{id:"h2o/h2o-danube3-500m",name:"H2O Danube3 500M",provider:"H2O.ai",family:"H2O Danube",version:"3",release_date:"2024-10-01",architecture:{type:"dense",total_params_b:.5,active_params_b:.5,num_layers:24,hidden_dim:896,num_heads:14,num_kv_heads:2,head_dim:64,intermediate_dim:4864,vocab_size:32e3},context_window:8192,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:1,weights_fp8_gb:.5,weights_int4_gb:.25,kv_per_token_bytes:12288,activation_estimate_gb:.1},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","ollama"],api_pricing:{},tags:["chat","open-weight","lightweight","mobile","estimated-architecture"],huggingface_id:"h2oai/h2o-danube3-500m-chat",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/h2oai/h2o-danube3-500m-chat",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.181Z",license:"apache-2.0"},{id:"huggingface/kokoro-82m",name:"Kokoro TTS 82M",provider:"Hexagrad",family:"Kokoro",version:"1",release_date:"2025-01-15",architecture:{type:"dense",total_params_b:.082,active_params_b:.082,num_layers:6,hidden_dim:512,num_heads:8,num_kv_heads:8,head_dim:64,intermediate_dim:2048,vocab_size:32e3},context_window:2048,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:.164,weights_fp8_gb:.082,weights_int4_gb:.041,kv_per_token_bytes:6144,activation_estimate_gb:.01},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["ollama"],api_pricing:{},tags:["text-to-speech","open-weight","edge"],huggingface_id:"pratyaypandey/Kokoro-82M",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/pratyaypandey/Kokoro-82M",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-16T09:29:53.000Z"},{id:"huggingface/smollm-135m",name:"SmolLM 135M",provider:"Hugging Face",family:"SmolLM",ver

1sion:"1",release_date:"2024-07-17",architecture:{type:"dense",total_params_b:.135,active_params_b:.135,num_layers:30,hidden_dim:576,num_heads:9,num_kv_heads:3,head_dim:64,intermediate_dim:1536,vocab_size:49152},context_window:2048,supported_precisions:["bf16","fp8"],default_precision:"bf16",memory:{weights_bf16_gb:.27,weights_fp8_gb:.135,weights_int4_gb:.068,kv_per_token_bytes:11520,activation_estimate_gb:.03},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","tgi","ollama"],api_pricing:{},tags:["chat","open-weight","lightweight","edge"],huggingface_id:"HuggingFaceTB/SmolLM2-135M-Instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/HuggingFaceTB/SmolLM2-135M-Instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.181Z",license:"apache-2.0"},{id:"huggingface/smollm-360m",name:"SmolLM 360M",provider:"Hugging Face",family:"SmolLM",version:"1",release_date:"2024-07-17",architecture:{type:"dense",total_params_b:.36,active_params_b:.36,num_layers:32,hidden_dim:960,num_heads:15,num_kv_heads:5,head_dim:64,intermediate_dim:2560,vocab_size:49152},context_window:2048,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:.72,weights_fp8_gb:.36,weights_int4_gb:.18,kv_per_token_bytes:20480,activation_estimate_gb:.05},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","tgi","ollama"],api_pricing:{},tags:["chat","open-weight","lightweight","edge"],huggingface_id:"HuggingFaceTB/SmolLM-360M-Instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/HuggingFaceTB/SmolLM-360M-Instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.181Z",license:"apache-2.0"},{id:"huggingface/smollm2-1.7b",name:"SmolLM2 1.7B",provider:"Hugging Face",family:"SmolLM2",version:"2",release_date:"2024-11-01",architecture:{type:"dense",total_params_b:1.7,active_params_b:1.7,num_layers:24,hidden_dim:2048,num_heads:32,num_kv_heads:32,head_dim:64,intermediate_dim:8192,vocab_size:49152},context_window:8192,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:3.4,weights_fp8_gb:1.7,weights_int4_gb:.85,kv_per_token_bytes:196608,activation_estimate_gb:.3},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","ollama"],api_pricing:{},tags:["chat","open-weight","lightweight"],huggingface_id:"HuggingFaceTB/SmolLM2-1.7B-Instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/HuggingFaceTB/SmolLM2-1.7B-Instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.181Z",license:"apache-2.0"},{id:"huggingface/zephyr-7b",name:"Zephyr 7B",provider:"Hugging Face",family:"Zephyr",version:"beta",release_date:"2023-10-26",architecture:{type:"dense",total_params_b:7,active_params_b:7,num_layers:32,hidden_dim:4096,num_heads:32,num_kv_heads:8,head_dim:128,intermediate_dim:14336,vocab_size:32e3},context_window:32768,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:14,weights_fp8_gb:7,weights_int4_gb:3.5,kv_per_token_bytes:131072,activation_estimate_gb:1},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!0,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{},tags:["chat","open-weight"],huggingface_id:"HuggingFaceH4/zephyr-7b-alpha",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/HuggingFaceH4/zephyr-7b-alpha",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.181Z",license:"mit"},{id:"huggingfacem4/siglip-so400m-14-384-flash-attn2-navit",name:"siglip-so400m-14-384-flash-attn2-navit",tags:["transformers","safetensors","siglip","zero-shot-image-classification","custom_code","endpoints_compatible"],family:"HuggingFaceM4",memory:{weights_fp8_gb:.9,weights_bf16_gb:1.8,weights_int4_gb:.45,kv_per_token_bytes:124416,activation_estimate_gb:124416e-9},version:"14",provider:"HuggingFaceM4",is_partial:!1,api_pricing:{},architecture:{type:"dense",head_dim:72,num_heads:16,hidden_dim:1152,num_layers:27,vocab_size:32e3,num_kv_heads:16,total_params_b:.9,active_params_b:.9,intermediate_dim:4304},capabilities:{code:!1,math:!1,tools:!1,vision:!0,reasoning:!1,multilingual:!1,structured_output:!1},release_date:"2026-01-01",context_window:8192,huggingface_id:"HuggingFaceM4/siglip-so400m-14-384-flash-attn2-navit",default_precision:"bf16",supp

1orted_frameworks:["vllm"],supported_precisions:["bf16"],status:"active"},{id:"ibm-granite/granite-embedding-311m-multilingual-r2",name:"granite-embedding-311m-multilingual-r2",tags:[],family:"ibm-granite",memory:{weights_fp8_gb:7,weights_bf16_gb:14,weights_int4_gb:3.5,kv_per_token_bytes:131072,activation_estimate_gb:524288e-9},version:"311",provider:"ibm-granite",api_pricing:{},architecture:{type:"dense",head_dim:128,num_heads:32,hidden_dim:4096,num_layers:32,vocab_size:32e3,num_kv_heads:8,total_params_b:7,active_params_b:7,intermediate_dim:11008},capabilities:{code:!1,math:!1,tools:!1,vision:!1,reasoning:!1,multilingual:!0,structured_output:!1},release_date:"2026-04-20",context_window:8192,huggingface_id:"ibm-granite/granite-embedding-311m-multilingual-r2",default_precision:"bf16",supported_frameworks:["vllm"],supported_precisions:["bf16"],status:"draft",license:"apache-2.0",last_verified_at:"2026-06-25T15:18:14.181Z"},{id:"internlm/internlm2.5-20b",name:"InternLM 2.5 20B",provider:"Shanghai AI Lab",family:"InternLM 2.5",version:"2.5",release_date:"2024-07-03",architecture:{type:"dense",total_params_b:19.9,active_params_b:19.9,num_layers:48,hidden_dim:6144,num_heads:48,num_kv_heads:8,head_dim:128,intermediate_dim:16384,vocab_size:92544},context_window:262144,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:39.8,weights_fp8_gb:19.9,weights_int4_gb:9.95,kv_per_token_bytes:196608,activation_estimate_gb:1.5},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm","sglang","tgi"],api_pricing:{internlm:{input_per_m:.5,output_per_m:.5}},tags:["chat","open-weight","instruction-following"],huggingface_id:"internlm/internlm2_5-20b",huggingface_id_status:"manual",release_date_source:"https://huggingface.co/internlm/internlm2_5-20b",last_verified_at:"2026-06-25T16:19:04.533Z",license:"apache-2.0"},{id:"internlm/internlm2.5-7b",name:"InternLM 2.5 7B",provider:"Shanghai AI Lab",family:"InternLM 2.5",version:"2.5",release_date:"2024-07-03",architecture:{type:"dense",total_params_b:7.74,active_params_b:7.74,num_layers:32,hidden_dim:4096,num_heads:32,num_kv_heads:8,head_dim:128,intermediate_dim:14336,vocab_size:92544},context_window:1048576,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:15.48,weights_fp8_gb:7.74,weights_int4_gb:3.87,kv_per_token_bytes:131072,activation_estimate_gb:1},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm","sglang","tgi","ollama"],api_pricing:{internlm:{input_per_m:.2,output_per_m:.2}},tags:["chat","open-weight","instruction-following","long-context"],huggingface_id:"zzfive/ComfyChat-InternLM2.5-7b-v2-1",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/zzfive/ComfyChat-InternLM2.5-7b-v2-1",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-16T09:29:53.000Z"},{id:"internlm/internlm3-8b",name:"InternLM3 8B",provider:"Shanghai AI Lab",family:"InternLM",version:"3",release_date:"2025-01-15",architecture:{type:"dense",total_params_b:8,active_params_b:8,num_layers:32,hidden_dim:4096,num_heads:32,num_kv_heads:8,head_dim:128,intermediate_dim:14336,vocab_size:103168},context_window:32768,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:16,weights_fp8_gb:8,weights_int4_gb:4,kv_per_token_bytes:65536,activation_estimate_gb:.5},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm","sglang","tgi","llama-cpp"],api_pricing:{},tags:["chat","code","reasoning","open-weight"],huggingface_id:"internlm/internlm3-8b-instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/internlm/internlm3-8b-instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.181Z",license:"apache-2.0"},{id:"internlm/internvl2-26b",name:"InternVL2 26B",provider:"InternLM",family:"InternVL2",version:"2",release_date:"2024-0

17-15",architecture:{type:"dense",total_params_b:26,active_params_b:26,num_layers:48,hidden_dim:6144,num_heads:48,num_kv_heads:8,head_dim:128,intermediate_dim:16384,vocab_size:92544},context_window:32768,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:52,weights_fp8_gb:26,weights_int4_gb:13,kv_per_token_bytes:196608,activation_estimate_gb:1.5},capabilities:{tools:!1,vision:!0,structured_output:!1,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi"],api_pricing:{},tags:["chat","vision","vision-language","open-weight"],huggingface_id:"OpenGVLab/InternVL2-26B",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/OpenGVLab/InternVL2-26B",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.181Z",license:"mit"},{id:"intfloat/e5-mistral-7b-instruct",name:"e5-mistral-7b-instruct",provider:"intfloat",family:"intfloat",release_date:"2023-12-20",supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",supported_frameworks:["vllm","sglang","tgi","tensorrt-llm"],tags:["embedding","retrieval","rag","text-embedding","open-weight"],huggingface_id:"intfloat/e5-mistral-7b-instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/intfloat/e5-mistral-7b-instruct",release_date_verified_at:"2026-06-24T04:04:52.914Z",last_verified_at:"2026-06-24T04:04:52.914Z",license:"mit",modality:"text-embedding",specialized_architecture:{type:"encoder",total_params_b:7.11,notes:"E5-Mistral text embedding model: a Mistral-7B backbone (MistralModel) fine-tuned as a dense text encoder producing sentence/passage embeddings. 7.11B parameters. One of the strongest open-weight embedding models on MTEB-style retrieval benchmarks."},specialized_memory:{weights_bf16_gb:14.22,activation_estimate_gb:2},specialized_pricing:{kind:"free_open_weights"},description:"E5-Mistral-7B-Instruct is intfloat's open-weight text embedding model built on a Mistral-7B backbone. It produces high-quality dense embeddings for retrieval, RAG and semantic search, and is a strong open baseline on MTEB."},{id:"intfloat/llm-retriever-base",name:"llm-retriever-base",family:"intfloat",license:"mit",provider:"intfloat",release_date:"2023-07-14",last_verified_at:"2026-06-22T12:35:40.483Z",release_date_source:"https://huggingface.co/intfloat/llm-retriever-base",release_date_verified_at:"2026-06-22T12:35:40.483Z",modality:"text-embedding",specialized_architecture:{type:"encoder",total_params_b:.109,notes:"LLM-Retriever base: a BERT-base-scale (0.11B) dense retriever that embeds queries and passages into a shared vector space for in-context example / document retrieval. Open-weight, transformers library."},specialized_memory:{weights_bf16_gb:.22,activation_estimate_gb:.1},specialized_pricing:{kind:"free_open_weights"},description:"LLM-Retriever-Base is intfloat's compact (0.11B) open-weight dense retriever for embedding queries and passages into a shared space — designed for retrieving in-context examples and documents to feed LLMs.",tags:["embedding","retrieval","dense-retriever","open-weight"]},{id:"intfloat/mme5-mllama-11b-instruct",name:"mmE5-mllama-11b-instruct",tags:["transformers","safetensors","mllama","image-text-to-text","mmeb","sentence-transformers","zero-shot-image-classification","en"],family:"intfloat",memory:{weights_fp8_gb:10.6,weights_bf16_gb:21.2,weights_int4_gb:5.3,kv_per_token_bytes:163840,activation_estimate_gb:65536e-8},version:"1.0",provider:"intfloat",api_pricing:{},architecture:{type:"dense",head_dim:128,num_heads:32,hidden_dim:4096,num_layers:40,vocab_size:128256,num_kv_heads:8,total_params_b:10.6,active_params_b:10.6,intermediate_dim:14336},capabilities:{code:!1,math:!1,tools:!1,vision:!0,reasoning:!1,multilingual:!1,structured_output:!1},release_date:"2025-02-13",context_window:131072,huggingface_id:"intfloat/mmE5-mllama-11b-instruct",default_precision:"bf16",supported_frameworks:["vllm"],supported_precisions:["bf16"],status:"active",license:"mit",last_verified_at:"2026-06-25T15:18:14.181Z"},{id:"intfloat/multilingual-e5-large-instruct",name:"multilingual-e5-large-instruct",tags:["sentence-transformers","onnx","safetensors","xlm-roberta","feature-extraction","mteb","transformers","multilingual"],family:"intfloat",memory:{weights_fp8_gb:.6,weights_bf16_gb:1.2,weights_int4_gb:.3,kv_per_token_bytes:98304,activation_estimate_gb:98304e-9},version:"1.0",provider:"intfloat",api_pricing:{},architecture:{type:"dense",head_dim:64,num_heads:16,hidden_dim:1024,num_layers:24,vocab_size:250002,num_kv_heads:16,total_params_b:.6,active_params_b:.6,intermediate_dim:4096},capabilities:{code:!1,math:!1,tools:!1,vision:!1,reasoning:!1,multilingual:!1,structured_output:!1},release_date:"2024-02-08",context_window:514,huggingface_id:"intfloat/multilingual-e5-large-instruct",default_precision:"bf16",supp

1orted_frameworks:["vllm"],supported_precisions:["bf16"],status:"active",license:"mit",last_verified_at:"2026-06-25T15:18:14.181Z"},{id:"intfloat/multilingual-e5-large",name:"Multilingual E5 Large",provider:"Microsoft",family:"E5",version:"1",release_date:"2024-02-08",architecture:{type:"dense",total_params_b:.56,active_params_b:.56,num_layers:24,hidden_dim:1024,num_heads:16,num_kv_heads:16,head_dim:64,intermediate_dim:4096,vocab_size:250002},context_window:512,supported_precisions:["bf16","fp8"],default_precision:"bf16",memory:{weights_bf16_gb:1.12,weights_fp8_gb:.56,weights_int4_gb:.28,kv_per_token_bytes:49152,activation_estimate_gb:.1},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","tgi"],api_pricing:{},tags:["embedding","multilingual","open-weight"],huggingface_id:"intfloat/multilingual-e5-large",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/intfloat/multilingual-e5-large",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.181Z",license:"mit"},{id:"japanese/elyza-13b",name:"ELYZA 13B",provider:"ELYZA",family:"ELYZA",version:"1",release_date:"2023-12-19",architecture:{type:"dense",total_params_b:13,active_params_b:13,num_layers:40,hidden_dim:5120,num_heads:40,num_kv_heads:40,head_dim:128,intermediate_dim:13824,vocab_size:32e3},context_window:4096,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:26,weights_fp8_gb:13,weights_int4_gb:6.5,kv_per_token_bytes:409600,activation_estimate_gb:1},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","tgi","ollama"],api_pricing:{},tags:["japanese","multilingual","open-weight"],huggingface_id:"elyza/ELYZA-japanese-Llama-2-13b-instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/elyza/ELYZA-japanese-Llama-2-13b-instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.181Z",license:"llama2"},{id:"japanese/japanese-stablelm-70b",name:"Japanese StableLM 70B",provider:"Stability AI",family:"StableLM",version:"70b-jp",release_date:"2024-03-28",architecture:{type:"dense",total_params_b:70,active_params_b:70,num_layers:80,hidden_dim:8192,num_heads:64,num_kv_heads:8,head_dim:128,intermediate_dim:28672,vocab_size:65536},context_window:8192,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:140,weights_fp8_gb:70,weights_int4_gb:35,kv_per_token_bytes:655360,activation_estimate_gb:3},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi"],api_pricing:{},tags:["japanese","multilingual","open-weight"],huggingface_id:"stabilityai/japanese-stablelm-instruct-beta-70b",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/stabilityai/japanese-stablelm-instruct-beta-70b",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.181Z",license:"llama2"},{id:"jina/jina-embeddings-v3",name:"Jina Embeddings v3",provider:"Jina AI",family:"Jina Embeddings",version:"3",release_date:"2024-09-18",architecture:{type:"dense",total_params_b:.57,active_params_b:.57,num_layers:24,hidden_dim:1024,num_heads:16,num_kv_heads:16,head_dim:64,intermediate_dim:4096,vocab_size:250002},context_window:8192,supported_precisions:["bf16","fp8"],default_precision:"bf16",memory:{weights_bf16_gb:1.14,weights_fp8_gb:.57,weights_int4_gb:.285,kv_per_token_bytes:98304,activation_estimate_gb:.1},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["tgi","tensorrt-llm"],api_pricing:{jina:{input_per_m:.008,output_per_m:.008}},tags:["embedding","multilingual","open-weight"],huggingface_id:"jinaai/jina-embeddings-v3",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/jinaai/jina-embeddings-v3",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.182Z",license:"cc-by-nc-4.0"},{id:"kaist/prometheus-2-7b",name:"Prometheus 2 7B",provider:"KAIST",family:"Prometheus",version:"2",release_date:"2024-05-02",architecture:{type:"dense",total_params_b:7.24,active_params_b:7.24,num_layers:32,hidden_dim:4096,num_heads:32,num_kv_heads:8,head_dim:128,intermediate_dim:14336,vocab_size:32e3},context_window:8192,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:14.48,weights_fp8_gb:7.24,weights_int4_gb:3.62,kv_per_token_bytes:131072,activation_estimate_gb:.8},capabilities:{tools:!1,vision:!1,structured_output:!0,code:!1,math:!1,multilingual:!1,reasoning:!0},supported_frameworks:["vllm","sglang","tgi"],api_pricing:{},tags:["chat","evaluation","judge","open-weight"],huggingface_id:"yeirr/prometheus-7b-v2-q5-k-m",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/yeirr/

1prometheus-7b-v2-q5-k-m",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-16T09:29:53.000Z"},{id:"korean/kullm-12.8b",name:"KULLM 12.8B",provider:"Korea University",family:"KULLM",version:"1",release_date:"2023-08-01",architecture:{type:"dense",total_params_b:12.8,active_params_b:12.8,num_layers:40,hidden_dim:5120,num_heads:40,num_kv_heads:40,head_dim:128,intermediate_dim:13824,vocab_size:46e3},context_window:4096,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:25.6,weights_fp8_gb:12.8,weights_int4_gb:6.4,kv_per_token_bytes:409600,activation_estimate_gb:1},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","tgi"],api_pricing:{},tags:["korean","multilingual","open-weight"],huggingface_id:"nlpai-lab/kullm-polyglot-12.8b-v2",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/nlpai-lab/kullm-polyglot-12.8b-v2",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.182Z",license:"apache-2.0"},{id:"legal/saullm-7b",name:"SaulLM 7B",provider:"Equall.ai",family:"SaulLM",version:"1",release_date:"2024-03-01",architecture:{type:"dense",total_params_b:7.2,active_params_b:7.2,num_layers:32,hidden_dim:4096,num_heads:32,num_kv_heads:8,head_dim:128,intermediate_dim:14336,vocab_size:32e3},context_window:8192,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:14.4,weights_fp8_gb:7.2,weights_int4_gb:3.6,kv_per_token_bytes:131072,activation_estimate_gb:.8},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","ollama"],api_pricing:{},tags:["legal","domain-specific","open-weight"],huggingface_id:"auslawbench/Cite-SaulLM-7B",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/auslawbench/Cite-SaulLM-7B",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.182Z",license:"cc-by-4.0"},{id:"lmsys/vicuna-13b",name:"Vicuna 13B",provider:"LMSYS",family:"Vicuna",version:"1.5",release_date:"2023-07-29",architecture:{type:"dense",total_params_b:13,active_params_b:13,num_layers:40,hidden_dim:5120,num_heads:40,num_kv_heads:40,head_dim:128,intermediate_dim:13824,vocab_size:32e3},context_window:4096,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:26,weights_fp8_gb:13,weights_int4_gb:6.5,kv_per_token_bytes:819200,activation_estimate_gb:1.5},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","ollama"],api_pricing:{},tags:["chat","open-weight"],huggingface_id:"jeffwan/vicuna-13b",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/jeffwan/vicuna-13b",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.182Z",license:"other"},{id:"lmsys/vicuna-33b",name:"Vicuna 33B",provider:"LMSYS",family:"Vicuna",version:"1.3",release_date:"2023-06-22",architecture:{type:"dense",total_params_b:33,active_params_b:33,num_layers:60,hidden_dim:6656,num_heads:52,num_kv_heads:52,head_dim:128,intermediate_dim:17920,vocab_size:32e3},context_window:2048,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:66,weights_fp8_gb:33,weights_int4_gb:16.5,kv_per_token_bytes:1597440,activation_estimate_gb:2},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","ollama"],api_pricing:{},tags:["chat","open-weight"],huggingface_id:"lmsys/vicuna-33b-v1.3",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/lmsys/vicuna-33b-v1.3",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-16T09:29:53.000Z"},{id:"lmsys/vicuna-7b",name:"Vicuna 7B",provider:"LMSYS",family:"Vicuna",version:"1.5",release_date:"2023-07-29",architecture:{type:"dense",total_params_b:7,active_params_b:7,num_layers:32,hidden_dim:4096,num_heads:32,num_kv_heads:32,head_dim:128,intermediate_dim:11008,vocab_size:32e3},context_window:4096,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:14,weights_fp8_gb:7,weights_int4_gb:3.5,kv_per_token_bytes:524288,activation_estimate_gb:1},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","ollama"],api_pricing:{},tags:["chat","open-weight"],huggingface_id:"AlekseyKorshuk/vicuna-7b",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/AlekseyKorshuk/vicuna-7b",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.182Z",license:"other"},{id:"medical/biomistral-7b",name:"BioMistral 7B",provider:"BioMistral",family:"BioMistral",version:"1",release_date:"2024-02-15",architecture:{type:"dense",total_params_b:7.2,active_params_b:7.2,num_layers:32,hidden_dim:4096,num_heads:32,num_kv_heads:8,head_dim:128,intermediate_dim:14336,vocab_size:32e3},context_window:32768,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:14.4,weights_fp8_gb:7.2,weights_int4_gb:3.6,kv_per_token_bytes:131072,activation_estimate_gb:.8},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","ollama"],api_pricing:{},tags:["medical","biomedical","domain-specific","open-weight"],huggingface_id:"BioMistral/BioMistral-7B",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/BioMistral/BioMistral-7B",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.182Z",license:"apache-2.0"},{id:"medical/meditron-70b",name:"Meditron 70B",provider:"EPFL",family:"Meditron",version:"1",release_date:"2023-11-27",architecture:{type:"dense",total_params_b:70,active_params_b:70,num_layers:80,hidden_dim:8192,num_heads:64,num_kv_heads:8,head_dim:128,intermediate_dim:28672,vocab_size:32e3},context_window:4096,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:140,weights_fp8_gb:70,weights_int4_gb:35,kv_per_token_bytes:655360,activation_estimate_gb:3},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm"],api_pricing:{},tags:["medical","domain-specific","open-weight"],huggingface_id:"epfl-llm/meditron-70b",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/epfl-llm/meditron-70b",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.182Z",license:"llama2"},{id:"meta-llama/codellama-13b",name:"Code Llama 13B",provider:"Meta",family:"Code Llama",version:"1",release_date:"2023-08-24",architecture:{type:"dense",total_params_b:13,active_params_b:13,num_layers:40,hidden_dim:5120,num_heads:40,num_kv_heads:40,head_dim:128,intermediate_dim:13824,vocab_size:32016},context_window:16384,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:26,weights_fp8_gb:13,weights_int4_gb:6.5,kv_per_token_bytes:819200,activation_estimate_gb:1.5},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!0,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{together:{input_per_m:.22,output_per_m:.22}},tags:["code","open-weight","code-specialized"],release_date_source:"https://www.llama.com/",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-16T09:29:53.000Z"},{id:"meta-llama/codellama-34b",name:"Code Llama 34B",provider:"Meta",family:"Code Llama",version:"1",release_date:"2023-08-24",architecture:{type:"dense",total_params_b:34,active_params_b:34,num_layers:48,hidden_dim:8192,num_heads:64,num_kv_heads:8,head_dim:128,intermediate_dim:22016,vocab_size:32016},context_window:1e5,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:68,weights_fp8_gb:34,weights_int4_gb:17,kv_per_token_bytes:196608,activation_estimate_gb:2},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!0,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{together:{input_per_m:.78,output_per_m:.78}},tags:["code","open-weight","code-specialized"],release_date_source:"https://www.llama.com/",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-16T09:29:53.000Z"},{id:"meta-llama/codellama-7b",name:"Code Llama 7B",provider:"Meta",family:"Code Llama",version:"1",release_date:"2023-08-24",architecture:{type:"dense",total_params_b:7,active_params_b:7,num_layers:32,hidden_dim:4096,num_heads:32,num_kv_heads:32,head_dim:128,intermediate_dim:11008,vocab_size:32016},context_window:16384,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:14,weights_fp8_gb:7,weights_int4_gb:3.5,kv_per_token_bytes:524288,activation_estimate_gb:1},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!0,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{together:{input_per_m:.2,output_per_m:.2},openrouter:{input_per_m:.7999999999999999,output_per_m:1.2}},tags:["code","open-weight","code-specialized"],huggingface_id:"meta-llama/CodeLlama-7b-Python-hf",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/meta-llama/CodeLlama-7b-Python-hf",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.182Z",license:"llama2"},{id:"meta-llama/llama-2-13b",name:"Llama 2 13B",provider:"Meta",family:"Llama 2",version:"2",release_date:"2023-07-18",architecture:{type:"dense",total_params_b:13,active_params_b:13,num_layers:40,hidden_dim:5120,num_heads:40,num_kv_heads:40,head_dim:128,intermediate_dim:13824,vocab_size:32e3},context_window:4096,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:26,weights_fp8_gb:13,weights_int4_gb:6.5,kv_per_token_bytes:819200,activation_estimate_gb:1},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{},tags:["chat","open-weight"],release_date_source:"https://www.llama.com/",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-16T09:29:53.000Z"},{id:"meta-llama/llama-2-70b",name:"Llama 2 70B",provider:"Meta",family:"Llama 2",version:"2",release_date:"2023-07-18",architecture:{type:"dense",total_params_b:70,active_params_b:70,num_layers:80,hidden_dim:8192,num_heads:64,num_kv_heads:64,head_dim:128,intermediate_dim:28672,vocab_size:32e3},context_window:4096,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:140,weights_fp8_gb:70,weights_int4_gb:35,kv_per_token_bytes:2621440,activation_estimate_gb:2.5},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm"],api_pricing:{together:{input_per_m:.9,output_per_m:.9}},tags:["chat","open-weight"],huggingface_id:"meta-llama/Llama-2-70b",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/meta-llama/Llama-2-70b",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.182Z",license:"llama2"},{id:"meta-llama/llama-2-7b",name:"Llama 2 7B",provider:"Meta",family:"Llama 2",version:"2",release_date:"2023-07-18",architecture:{type:"dense",total_params_b:7,active_params_b:7,num_layers:32,hidden_dim:4096,num_heads:32,num_kv_heads:32,head_dim:128,intermediate_dim:11008,vocab_size:32e3},context_window:4096,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:14,weights_fp8_gb:7,weights_int4_gb:3.5,kv_per_token_bytes:524288,activation_estimate_gb:.8},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{},tags:["chat","open-weight"],huggingface_id:"meta-llama/Llama-2-7b",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/meta-llama/Llama-2-7b",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.182Z",license:"llama2"},{id:"meta-llama/llama-3-70b",name:"Llama 3 70B",provider:"Meta",family:"Llama 3",version:"3",release_date:"2024-04-18",architecture:{type:"dense",total_params_b:70.6,active_params_b:70.6,num_layers:80,hidden_dim:8192,num_heads:64,num_kv_heads:8,head_dim:128,intermediate_dim:28672,vocab_size:128256},context_window:8192,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:14

11.2,weights_fp8_gb:70.6,weights_int4_gb:35.3,kv_per_token_bytes:327680,activation_estimate_gb:2.5},capabilities:{tools:!1,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm"],api_pricing:{together:{input_per_m:.88,output_per_m:.88},novita:{input_per_m:.51,output_per_m:.74},featherless:{input_per_m:0,output_per_m:0},openrouter:{input_per_m:.51,output_per_m:.74}},tags:["chat","code","open-weight"],huggingface_id:"meta-llama/Meta-Llama-3-70B-Instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/meta-llama/Meta-Llama-3-70B-Instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.182Z",license:"llama3"},{id:"meta-llama/llama-3-8b",name:"Llama 3 8B",provider:"Meta",family:"Llama 3",version:"3",release_date:"2024-04-18",architecture:{type:"dense",total_params_b:8,active_params_b:8,num_layers:32,hidden_dim:4096,num_heads:32,num_kv_heads:8,head_dim:128,intermediate_dim:14336,vocab_size:128256},context_window:8192,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:16,weights_fp8_gb:8,weights_int4_gb:4,kv_per_token_bytes:131072,activation_estimate_gb:1},capabilities:{tools:!1,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{together:{input_per_m:.2,output_per_m:.2},openrouter:{input_per_m:.04,output_per_m:.04}},tags:["chat","code","open-weight"],huggingface_id:"meta-llama/Meta-Llama-3-8B",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/meta-llama/Meta-Llama-3-8B",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.182Z",license:"llama3"},{id:"meta-llama/llama-3.1-405b",name:"Llama 3.1 405B",provider:"Meta",family:"Llama 3.1",version:"3.1",release_date:"2024-07-23",architecture:{type:"dense",total_params_b:405,active_params_b:405,num_layers:126,hidden_dim:16384,num_heads:128,num_kv_heads:8,head_dim:128,intermediate_dim:53248,vocab_size:128256},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:810,weights_fp8_gb:405,weights_int4_gb:202.5,kv_per_token_bytes:516096,activation_estimate_gb:5},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm"],api_pricing:{together:{input_per_m:3.5,output_per_m:3.5},fireworks:{input_per_m:3,output_per_m:3},openrouter:{input_per_m:1,output_per_m:1}},tags:["chat","code","instruction-following","open-weight","flagship"],huggingface_id:"meta-llama/Llama-3.1-405B-Instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/meta-llama/Llama-3.1-405B-Instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.182Z",license:"llama3.1"},{id:"meta-llama/llama-3.1-70b",name:"Llama 3.1 70B",provider:"Meta",family:"Llama 3.1",version:"3.1",release_date:"2024-07-23",architecture:{type:"dense",total_params_b:70.6,active_params_b:70.6,num_layers:80,hidden_dim:8192,num_heads:64,num_kv_heads:8,head_dim:128,intermediate_dim:28672,vocab_size:128256},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:141.2,weights_fp8_gb:70.6,weights_int4_gb:35.3,kv_per_token_bytes:327680,activation_estimate_gb:2.5},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm"],api_pricing:{together:{input_per_m:.88,output_per_m:.88},fireworks:{input_per_m:.9,output_per_m:.9},groq:{input_per_m:.59,output_per_m:.79},scaleway:{input_per_m:0,output_per_m:0},featherless:{input_per_m:0,output_per_m:0},openrouter:{input_per_m:.59,output_per_m:.79},"together-ai":{input_per_m:.88,output_per_m:.88},deepinfra:{input_per_m:.35,output_per_m:.4}},tags:["chat","code","instruction-following","open-weight"],huggingface_id:"meta-llama/Llama-3.1-70B-Instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/meta-llama/Llama-3.1-70B-Instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.182Z",license:"llama3.1"}

1,{id:"meta-llama/llama-3.1-8b",name:"Llama 3.1 8B",provider:"Meta",family:"Llama 3.1",version:"3.1",release_date:"2024-07-23",architecture:{type:"dense",total_params_b:8.03,active_params_b:8.03,num_layers:32,hidden_dim:4096,num_heads:32,num_kv_heads:8,head_dim:128,intermediate_dim:14336,vocab_size:128256},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:16.06,weights_fp8_gb:8.03,weights_int4_gb:4.015,kv_per_token_bytes:131072,activation_estimate_gb:1},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{together:{input_per_m:.18,output_per_m:.18},fireworks:{input_per_m:.2,output_per_m:.2},groq:{input_per_m:.05,output_per_m:.08},novita:{input_per_m:.02,output_per_m:.05},cerebras:{input_per_m:.1,output_per_m:.1},scaleway:{input_per_m:0,output_per_m:0},featherless:{input_per_m:0,output_per_m:0},openrouter:{input_per_m:.02,output_per_m:.05},"together-ai":{input_per_m:.18,output_per_m:.18},deepinfra:{input_per_m:.05,output_per_m:.08}},tags:["chat","code","instruction-following","open-weight"],huggingface_id:"meta-llama/Llama-3.1-8B-Instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/meta-llama/Llama-3.1-8B-Instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.183Z",license:"llama3.1"},{id:"meta-llama/llama-3.2-11b-vision",name:"Llama 3.2 11B Vision",provider:"Meta",family:"Llama 3.2",version:"3.2",release_date:"2024-09-25",architecture:{type:"dense",total_params_b:11,active_params_b:11,num_layers:40,hidden_dim:4096,num_heads:32,num_kv_heads:8,head_dim:128,intermediate_dim:14336,vocab_size:128256},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:22,weights_fp8_gb:11,weights_int4_gb:5.5,kv_per_token_bytes:163840,activation_estimate_gb:1},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{together:{input_per_m:.18,output_per_m:.18},fireworks:{input_per_m:.2,output_per_m:.2},openrouter:{input_per_m:.245,output_per_m:.245}},tags:["chat","vision","vision-language","open-weight"],huggingface_id:"meta-llama/Llama-3.2-11B-Vision-Instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/meta-llama/Llama-3.2-11B-Vision-Instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.183Z",license:"llama3.2"},{id:"meta-llama/llama-3.2-1b",name:"Llama 3.2 1B",provider:"Meta",family:"Llama 3.2",version:"3.2",release_date:"2024-09-25",architecture:{type:"dense",total_params_b:1.24,active_params_b:1.24,num_layers:16,hidden_dim:2048,num_heads:32,num_kv_heads:8,head_dim:64,intermediate_dim:8192,vocab_size:128256},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:2.48,weights_fp8_gb:1.24,weights_int4_gb:.62,kv_per_token_bytes:32768,activation_estimate_gb:.3},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{together:{input_per_m:.03,output_per_m:.03},fireworks:{input_per_m:.1,output_per_m:.1},novita:{input_per_m:0,output_per_m:0},featherless:{input_per_m:0,output_per_m:0},openrouter:{input_per_m:.027,output_per_m:.2}},tags:["chat","code","instruction-following","open-weight","lightweight"],huggingface_id:"meta-llama/Llama-3.2-1B-Instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/meta-llama/Llama-3.2-1B-Instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.183Z",license:"llama3.2"},{id:"meta-llama/llama-3.2-3b",name:"Llama 3.2 3B",provider:"Meta",family:"Llama 3.2",version:"3.2",release_date:"2024-09-25",architecture:{type:"dense",total_params_b:3.21,active_params_b:3.21,num_layers:28,hidden_dim:3072,num_heads:24,num_kv_heads:8,head_dim:128,intermediate_dim:8192,vocab_size:128256},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:6.42,weights_fp8_gb:3.21,weights_int4_gb:1.605,kv_per_token_bytes:114688,activation_estimate_gb:.5},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{together:{input_per_m:.06,output_per_m:.06},fireworks:{input_per_m:.1,output_per_m:.1},openrouter:{input_per_m:.051,output_per_m:.33999999999999997}},tags:["chat","code","instruction-following","open-weight","lightweight"],huggingface_id:"meta-llama/Llama-3.2-3B-Instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/meta-llama/Llama-3.2-3B-Instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.183Z",license:"llama3.2"},{id:"meta-llama/llama-3.2-90b-vision-instruct",name:"Llama 3.2 90B Vision Instru

1ct",provider:"Meta",family:"Llama 3.2",version:"3.2",release_date:"2024-09-25",architecture:{type:"dense",total_params_b:88.8,active_params_b:88.8,num_layers:80,hidden_dim:8192,num_heads:64,num_kv_heads:8,head_dim:128,intermediate_dim:28672,vocab_size:128256},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:177.6,weights_fp8_gb:88.8,weights_int4_gb:44.4,kv_per_token_bytes:655360,activation_estimate_gb:4},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm"],api_pricing:{together:{input_per_m:1.2,output_per_m:1.2}},tags:["chat","vision","multimodal","instruct","open-weight"],huggingface_id:"meta-llama/Llama-3.2-90B-Vision-Instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/meta-llama/Llama-3.2-90B-Vision-Instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.183Z",license:"llama3.2"},{id:"meta-llama/llama-3.2-90b-vision",name:"Llama 3.2 90B Vision",provider:"Meta",family:"Llama 3.2",version:"3.2",release_date:"2024-09-25",architecture:{type:"dense",total_params_b:90,active_params_b:90,num_layers:80,hidden_dim:8192,num_heads:64,num_kv_heads:8,head_dim:128,intermediate_dim:28672,vocab_size:128256},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:180,weights_fp8_gb:90,weights_int4_gb:45,kv_per_token_bytes:327680,activation_estimate_gb:3},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm"],api_pricing:{together:{input_per_m:1.2,output_per_m:1.2},fireworks:{input_per_m:.9,output_per_m:.9}},tags:["chat","vision","vision-language","open-weight"],huggingface_id:"meta-llama/Llama-3.2-90B-Vision-Instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/meta-llama/Llama-3.2-90B-Vision-Instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.183Z",license:"llama3.2"},{id:"meta-llama/llama-3.3-70b",name:"Llama 3.3 70B",provider:"Meta",family:"Llama 3.3",version:"3.3",release_date:"2024-12-06",architecture:{type:"dense",total_params_b:70.6,active_params_b:70.6,num_layers:80,hidden_dim:8192,num_heads:64,num_kv_heads:8,head_dim:128,intermediate_dim:28672,vocab_size:128256},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:141.2,weights_fp8_gb:70.6,weights_int4_gb:35.3,kv_per_token_bytes:327680,activation_estimate_gb:2.5},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{together:{input_per_m:.88,output_per_m:.88},fireworks:{input_per_m:.9,output_per_m:.9},groq:{input_per_m:.59,output_per_m:.79},novita:{input_per_m:.135,output_per_m:.4},hyperbolic:{input_per_m:.4,output_per_m:.4},sambanova:{input_per_m:.6,output_per_m:1.2},ovhcloud:{input_per_m:.74,output_per_m:.74},scaleway:{input_per_m:0,output_per_m:0},featherless:{input_per_m:0,output_per_m:0},openrouter:{input_per_m:.1,output_per_m:.32}},tags:["chat","code","instruction-following","open-weight"],huggingface_id:"meta-llama/Llama-3.3-70B-Instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/meta-llama/Llama-3.3-70B-Instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.183Z",license:"llama3.3"},{id:"meta-llama/llama-3.3-8b",name:"Llama 3.3 8B",provider:"Meta",family:"Llama 3.3",version:"3.3",release_date:"2025-01-15",architecture:{type:"dense",total_params_b:8,active_params_b:8,num_layers:32,hidden_dim:4096,num_heads:32,num_kv_heads:8,head_dim:128,intermediate_dim:14336,vocab_size:128256},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:16,weights_fp8_gb:8,weights_int4_gb:4,kv_per_token_bytes:131072,activation_estimate_gb:1},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{together:{input_per_m:.18,output_per_m:.18}},tags:["chat","code","open-weight","estimated-architecture"],release_date_source:"https://www.llama.com/",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-16T09:29:53.000Z"},{id:"meta-llama/llama-4-maverick-400b",name:"Llama 4 Maverick",provider:"Meta",family:"Llama 4",version:"4",release_date:"2025-04-05",architecture:{type:"moe",total_params_b:400,active_params_b:17,num_layers:96,hidden_dim:5120,num_heads:40,num_kv_heads:8,head_dim:128,intermediate_dim:13568,vocab_size:202048,num_experts:128,experts_per_token:1},context_window:1048576,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:800,weights_fp8_gb:400,weights_int4_gb:200,kv_per_token_bytes:393216,activation_estimate_gb:3},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tensorrt-llm"],api_pricing:{together:{input_per_m:1.2,output_per_m:1.8},fireworks:{input_per_m:1.5,output_per_m:2},sambanova:{input_per_m:.63,output_per_m:1.8},openrouter:{input_per_m:.15,output_per_m:.6}},tags:["chat","code","multimodal","moe","open-weight","flagship"],huggingface_id:"meta-llama/Llama-4-Maverick-17B-128E-Instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/meta-llama/Llama-4-Maverick-17B-128E-Instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.183Z",license:"other"},{id:"meta-llama/llama-4-maverick",name:"Llama 4 Maverick",provider:"Meta",family:"Llama 4",version:"Maverick",release_date:"2025-04-05",release_date_source:"https://huggingface.co/meta-llama/Llama-4-Maverick-17B-128E",architecture:{type:"moe",total_params_b:400,active_params_b:17,num_layers:48,hidden_dim:5120,num_heads:40,num_kv_heads:8,head_dim:128,intermediate_dim:8192,vocab_size:202048,num_experts:128,experts_per_token:1,shared_experts:1},context_window:1048576,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:745.

11,weights_fp8_gb:372.5,weights_int4_gb:186.3,kv_per_token_bytes:196608,activation_estimate_gb:.001},api_pricing:{openrouter:{input_per_m:.2,output_per_m:.7}},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang"],tags:["chat","code","moe","vision","multilingual","open-weight","flagship"],huggingface_id:"meta-llama/Llama-4-Maverick-17B-128E",huggingface_id_status:"auto_resolved",license:"other"},{id:"meta-llama/llama-4-scout-17b",name:"Llama 4 Scout",provider:"Meta",family:"Llama 4",version:"4",release_date:"2025-04-05",architecture:{type:"moe",total_params_b:109,active_params_b:17,num_layers:48,hidden_dim:5120,num_heads:40,num_kv_heads:8,head_dim:128,intermediate_dim:13568,vocab_size:202048,num_experts:16,experts_per_token:1},context_window:0xa00000,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:218,weights_fp8_gb:109,weights_int4_gb:54.5,kv_per_token_bytes:196608,activation_estimate_gb:2},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tensorrt-llm"],api_pricing:{together:{input_per_m:.18,output_per_m:.3},fireworks:{input_per_m:.2,output_per_m:.35},novita:{input_per_m:.18,output_per_m:.59},groq:{input_per_m:.11,output_per_m:.34},openrouter:{input_per_m:.08,output_per_m:.3}},tags:["chat","code","multimodal","moe","open-weight","long-context"],huggingface_id:"meta-llama/Llama-4-Scout-17B-16E-Instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/meta-llama/Llama-4-Scout-17B-16E-Instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.183Z",license:"other"},{id:"meta-llama/llama-4-scout",name:"Llama 4 Scout",provider:"Meta",family:"Llama 4",version:"Scout",release_date:"2025-04-05",release_date_source:"https://huggingface.co/meta-llama/Llama-4-Scout-17B-16E",architecture:{type:"moe",total_params_b:109,active_params_b:17,num_layers:48,hidden_dim:5120,num_heads:40,num_kv_heads:8,head_dim:128,intermediate_dim:8192,vocab_size:202048,num_experts:16,experts_per_token:1,shared_experts:1},context_window:0xa00000,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:203,weights_fp8_gb:101.5,weights_int4_gb:50.8,kv_per_token_bytes:196608,activation_estimate_gb:.001},api_pricing:{openrouter:{input_per_m:.1,output_per_m:.3}},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang"],tags:["chat","code","moe","vision","multilingual","open-weight"],huggingface_id:"meta-llama/Llama-4-Scout-17B-16E",huggingface_id_status:"auto_resolved",license:"other"},{id:"meta-llama/llama-guard-3-1b",name:"Llama Guard 3 1B",provider:"Meta",family:"Llama Guard",version:"3",release_date:"2024-12-06",architecture:{type:"dense",total_params_b:1,active_params_b:1,num_layers:16,hidden_dim:2048,num_heads:32,num_kv_heads:8,head_dim:64,intermediate_dim:8192,vocab_size:128256},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:2,weights_fp8_gb:1,weights_int4_gb:.5,kv_per_token_bytes:16384,activation_estimate_gb:.2},capabilities:{tools:!1,vision:!1,structured_output:!0,code:!1,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","ollama"],api_pricing:{},tags:["safety","classification","open-weight","lightweight"],huggingface_id:"meta-llama/Llama-Guard-3-1B",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/meta-llama/Llama-Guard-3-1B",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.183Z",license:"llama3.2"},{id:"meta-llama/llama-guard-3-8b",name:"Llama Guard 3 8B",provider:"Meta",family:"Llama Guard",version:"3",release_date:"2024-07-23",architecture:{type:"dense",total_params_b:8,active_params_b:8,num_layers:32,hidden_dim:4096,num_heads:32,num_kv_heads:8,head_dim:128,intermediate_dim:14336,vocab_size:128256},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:16,weights_fp8_gb:8,weights_int4_gb:4,kv_per_token_bytes:131072,activation_estimate_gb:1},capabilities:{tools:!1,vision:!1,structured_output:!0,code:!1,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{together:{input_per_m:.2,output_per_m:.2},openrouter:{input_per_m:.48,output_per_m:.03}},tags:["safety","classifier","moderation","open-weight"],huggingface_id:"meta-llama/Llama-Guard-3-8B",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/meta-llama/Llama-Guard-3-8B",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.183Z",license:"llama3.1"},{id:"meta/seamless-m4t-v2-large",name:"SeamlessM4T v2 Large",provider:"Meta",family:"SeamlessM4T",version:"2",release_date:"2023-12-01",architecture:{type:"dense",total_params_b:2.3,active_params_b:2.3,num_layers:24,hidden_dim:1024,num_heads:16,num_kv_heads:16,head_dim:64,intermediate_dim:4096,vocab_size:256102},context_window:4096,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:4.6,weights_fp8_gb:2.3,weights_int4_gb:1.15,kv_per_token_bytes:98304,activation_estimate_gb:.3},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["vllm"],api_pricing:{},tags:["audio","translation","multimodal","open-weight","estimated-architecture"],release_date_source:"https://www.llama.com/",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-16T09:29:53.000Z"},{id:"microsoft/florence-2-large",name:"Florence 2 Large",provider:"Microsoft",family:"Florence",version:"2",release_date:"2024-06-18",architecture:{type:"dense",total_params_b:.77,active_params_b:.77,num_layers:24,hidden_dim:1024,num_heads:16,num_kv_heads:16,head_dim:64,intermediate_dim:4096,vocab_size:51289},context_window:2048,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:1.54,weights_fp8_gb:.77,weights_int4_gb:.385,kv_per_token_bytes:98304,activation_estimate_gb:.2},capabilities:{tools:!1,vision:!0,structured_output:!0,code:!1,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm"],api_pricing:{},tags:["vision","multimodal","open-weight","lightweight","estimated-architecture"],huggingface_id:"microsoft/Florence-2-large",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/microsoft/Florence-2-large",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.183Z",license:"mit"},{id:"microsoft/orca-2-13b",name:"Orca 2 13B",provider:"Microsoft",family:"Orca",ver

1sion:"2",release_date:"2023-11-21",architecture:{type:"dense",total_params_b:13,active_params_b:13,num_layers:40,hidden_dim:5120,num_heads:40,num_kv_heads:40,head_dim:128,intermediate_dim:13824,vocab_size:32e3},context_window:4096,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:26,weights_fp8_gb:13,weights_int4_gb:6.5,kv_per_token_bytes:819200,activation_estimate_gb:1},capabilities:{tools:!1,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!1,reasoning:!0},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{},tags:["chat","reasoning","instruction-following","open-weight"],huggingface_id:"microsoft/Orca-2-13b",huggingface_id_status:"manual",release_date_source:"https://huggingface.co/microsoft/Orca-2-13b",last_verified_at:"2026-06-25T16:19:04.533Z",license:"other"},{id:"microsoft/phi-1.5",name:"Phi 1.5",provider:"Microsoft",family:"Phi",version:"1.5",release_date:"2023-09-11",architecture:{type:"dense",total_params_b:1.3,active_params_b:1.3,num_layers:24,hidden_dim:2048,num_heads:32,num_kv_heads:32,head_dim:64,intermediate_dim:8192,vocab_size:51200},context_window:2048,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:2.6,weights_fp8_gb:1.3,weights_int4_gb:.65,kv_per_token_bytes:196608,activation_estimate_gb:.2},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!0,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","ollama"],api_pricing:{},tags:["code","open-weight","lightweight"],huggingface_id:"microsoft/phi-1_5",huggingface_id_status:"manual",release_date_source:"https://huggingface.co/microsoft/phi-1_5",last_verified_at:"2026-06-25T16:19:04.533Z",license:"mit"},{id:"microsoft/phi-1",name:"Phi 1",provider:"Microsoft",family:"Phi",version:"1",release_date:"2023-06-20",architecture:{type:"dense",total_params_b:1.3,active_params_b:1.3,num_layers:24,hidden_dim:2048,num_heads:32,num_kv_heads:32,head_dim:64,intermediate_dim:8192,vocab_size:51200},context_window:2048,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:2.6,weights_fp8_gb:1.3,weights_int4_gb:.65,kv_per_token_bytes:98304,activation_estimate_gb:.2},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!0,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","tgi","ollama"],api_pricing:{},tags:["code","open-weight","lightweight"],huggingface_id:"microsoft/Phi-3-mini-128k-instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/microsoft/Phi-3-mini-128k-instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.183Z",license:"mit"},{id:"microsoft/phi-2",name:"Phi 2",provider:"Microsoft",family:"Phi",version:"2",release_date:"2023-12-13",architecture:{type:"dense",total_params_b:2.7,active_params_b:2.7,num_layers:32,hidden_dim:2560,num_heads:32,num_kv_heads:32,head_dim:80,intermediate_dim:10240,vocab_size:51200},context_window:2048,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:5.4,weights_fp8_gb:2.7,weights_int4_gb:1.35,kv_per_token_bytes:327680,activation_estimate_gb:.3},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!0,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{},tags:["chat","code","lightweight","edge","open-weight"],huggingface_id:"microsoft/phi-2",huggingface_id_status:"manual",release_date_source:"https://huggingface.co/microsoft/phi-2",last_verified_at:"2026-06-25T16:19:04.534Z",license:"mit"},{id:"microsoft/phi-3-medium-14b",name:"Phi 3 Medium 14B",provider:"Microsoft",family:"Phi 3",version:"3",release_date:"2024-05-21",architecture:{type:"dense",total_params_b:14,active_params_b:14,num_layers:40,hidden_dim:5120,num_heads:40,num_kv_heads:10,head_dim:128,intermediate_dim:17920,vocab_size:32064},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:28,weights_fp8_gb:14,weights_int4_gb:7,kv_per_token_bytes:204800,activation_estimate_gb:1.5},capabilities:{tools:!1,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{},tags:["chat","code","open-weight"],huggingface_id:"microsoft/Phi-3-medium-4k-instru

1ct",huggingface_id_status:"manual",release_date_source:"https://huggingface.co/microsoft/Phi-3-medium-4k-instruct",last_verified_at:"2026-06-25T16:19:04.534Z",license:"mit"},{id:"microsoft/phi-3-mini-3.8b",name:"Phi 3 Mini 3.8B",provider:"Microsoft",family:"Phi 3",version:"3",release_date:"2024-04-23",architecture:{type:"dense",total_params_b:3.8,active_params_b:3.8,num_layers:32,hidden_dim:3072,num_heads:32,num_kv_heads:32,head_dim:96,intermediate_dim:8192,vocab_size:32064},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:7.6,weights_fp8_gb:3.8,weights_int4_gb:1.9,kv_per_token_bytes:393216,activation_estimate_gb:.5},capabilities:{tools:!1,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{},tags:["chat","code","open-weight","lightweight"],huggingface_id:"microsoft/Phi-3-mini-4k-instruct",huggingface_id_status:"manual",release_date_source:"https://huggingface.co/microsoft/Phi-3-mini-4k-instruct",last_verified_at:"2026-06-25T16:19:04.534Z",license:"mit"},{id:"microsoft/phi-3-small-7b",name:"Phi 3 Small 7B",provider:"Microsoft",family:"Phi 3",version:"3",release_date:"2024-05-21",architecture:{type:"dense",total_params_b:7,active_params_b:7,num_layers:32,hidden_dim:4096,num_heads:32,num_kv_heads:8,head_dim:128,intermediate_dim:14336,vocab_size:100352},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:14,weights_fp8_gb:7,weights_int4_gb:3.5,kv_per_token_bytes:131072,activation_estimate_gb:1},capabilities:{tools:!1,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{},tags:["chat","code","open-weight"],huggingface_id:"microsoft/Phi-3-small-8k-instruct",huggingface_id_status:"manual",release_date_source:"https://huggingface.co/microsoft/Phi-3-small-8k-instruct",last_verified_at:"2026-06-25T16:19:04.534Z",license:"mit"},{id:"microsoft/phi-3.5-moe",name:"Phi 3.5 MoE",provider:"Microsoft",family:"Phi",version:"3.5-moe",release_date:"2024-08-20",architecture:{type:"moe",total_params_b:41.9,active_params_b:6.6,num_layers:32,hidden_dim:4096,num_heads:32,num_kv_heads:8,head_dim:128,intermediate_dim:6400,vocab_size:32064,num_experts:16,experts_per_token:2},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:83.8,weights_fp8_gb:41.9,weights_int4_gb:20.95,kv_per_token_bytes:131072,activation_estimate_gb:1},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm"],api_pricing:{deepinfra:{input_per_m:.1,output_per_m:.3}},tags:["chat","code","moe","open-weight"],huggingface_id:"microsoft/Phi-3.5-MoE-instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/microsoft/Phi-3.5-MoE-instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.184Z",license:"mit"},{id:"microsoft/phi-3.5-vision",name:"Phi 3.5 Vision",provider:"Microsoft",family:"Phi 3.5",version:"3.5",release_date:"2024-08-20",architecture:{type:"dense",total_params_b:4.2,active_params_b:4.2,num_layers:32,hidden_dim:3072,num_heads:32,num_kv_heads:32,head_dim:96,intermediate_dim:8192,vocab_size:32064},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:8.4,weights_fp8_gb:4.2,weights_int4_gb:2.1,kv_per_token_bytes:393216,activation_estimate_gb:.5},capabilities:{tools:!1,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{},tags:["chat","vision","vision-language","lightweight","open-weight"],huggingface_id:"microsoft/Phi-3.5-vision-instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/microsoft/Phi-3.5-vision-instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.184Z",license:"mit"},{id:"microsoft/phi-4-mini",name:"Phi 4 Mini",provider:"Microsoft",family:"Phi",version:"4-mini",release_date:"2025-02-27",architecture:{type:"dense",total_params_b:3.8,active_params_b:3.8,num_layers:32,hidden_dim:3072,num_heads:24,num_kv_heads:8,head_dim:128,intermediate_dim:8192,vocab_size:100352},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:7.6,weights_fp8_gb:3.8,weights_int4_gb:1.9,kv_per_token_bytes:65536,activation_estimate_gb:.5},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm","sglang","tgi","ollama"],api_pricing:{openrouter:{input_per_m:.08,output_per_m:.35}},tags:["chat","code","reasoning","open-weight","lightweight","estimated-architecture"],huggingface_id:"microsoft/Phi-4-mini-instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/microsoft/Phi-4-mini-instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.184Z",license:"mit"},{id:"microsoft/phi-4",name:"Phi-4",provider:"Microsoft",family:"Phi",version:"4",release_date:"2024-12-12",architecture:{type:"dense",total_params_b:14.7,active_params_b:14.7,num_layers:40,hidden_dim:5120,num_heads:40,num_kv_heads:10,head_dim:128,intermediate_dim:17920,vocab_size:100352},context_window:16384,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:29.4,weights_fp8_gb:14.7,weights_int4_gb:7.35,kv_per_token_bytes:204800,activation_estimate_gb:1.5},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{azure:{input_per_m:.07,output_per_m:.14},together:{input_per_m:.2,output_per_m:.2},openrouter:{input_per_m:.065,output_per_m:.14}},tags:["chat","code","math","reasoning","open-weight"],huggingface_id:"microsoft/Phi-4-mini-instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/microsoft/Phi-4-mini-instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.184Z",license:"mit"},{id:"microsoft/wizardmath-70b",name:"WizardMath 70B",provider:"Microsoft",family:"WizardMath",version:"1",release_date:"2023-08-11",architecture:{type:"dense",total_params_b:70,active_params_b:70,num_layers:80,hidden_dim:8192,num_heads:64,num_kv_heads:8,head_dim:128,intermediate_dim:28672,vocab_size:32e3},context_window:4096,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:140,weights_fp8_gb:70,weights_int4_gb:35,kv_per_token_bytes:327680,activation_estimate_gb:2.5},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!0,multilingual:!1,reasoning:!0},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm"],api_pricing:{},tags:["math","math-specialized","reasoning","open-weight"],huggingface_id:"WizardLMTeam/WizardMath-70B-V1.0",huggingface_id_status:"manual",release_date_source:"https://huggingface.co/WizardLMTeam/WizardMath-70B-V1.0",last_verified_at:"2026-06-25T16:19:04.534Z",license:"llama2"},{id:"minimax/minimax-m2-7",name:"MiniMax M2.7",provider:"MiniMax",family:"MiniMax M",version:"2.7",release_date:"2026-03-18",release_date_source:"https://platform.minimax.io/docs/release-notes/models",release_date_verified_at:"2026-05-13T14:42:10Z",last_verified_at:"2026-06-25T15:18:14.184Z",architecture:{type:"moe",total_params_b:456,active_params_b:45.9,num_layers:80,hidden_dim:6144,num_heads:48,num_kv_heads:8,head_dim:128,intermediate_dim:24576,vocab_size:200064},context_window:1048576,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:912,weights_fp8_gb:456,weights_int4_gb:228,kv_per_token_bytes:131072,activation_estimate_gb:4},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm","sglang"],api_pricing:{minimax:{input_per_m:.7,output_per_m:2.8},together:{input_per_m:.3,output_per_m:1.2},novita:{input_per_m:.3,output_per_m:1.2},fireworks:{input_per_m:0,output_per_m:0},featherless:{input_per_m:0,output_per_m:0}},tags:["chat","code","reasoning","vision","long-context","open-weight"],huggingface_id:"MiniMaxAI/MiniMax-M2.7",huggingface_id_status:"auto_resolved",license:"other"}

1,{id:"minimax/minimax-text-01",name:"MiniMax-Text-01",provider:"MiniMax",family:"MiniMax",version:"Text-01",release_date:"2025-01-15",architecture:{type:"moe",total_params_b:456,active_params_b:45.9,num_layers:80,hidden_dim:6144,num_heads:48,num_kv_heads:8,head_dim:128,intermediate_dim:16384,vocab_size:200064,num_experts:32,experts_per_token:2,shared_experts:0},context_window:1048576,supported_precisions:["bf16","fp8","int4"],default_precision:"fp8",memory:{weights_bf16_gb:912,weights_fp8_gb:456,weights_int4_gb:228,kv_per_token_bytes:163840,activation_estimate_gb:3},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm","sglang"],api_pricing:{minimax:{input_per_m:1,output_per_m:5}},tags:["chat","code","reasoning","moe","open-weight","flagship","long-context"],huggingface_id:"MiniMaxAI/MiniMax-Text-01",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/MiniMaxAI/MiniMax-Text-01",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-16T09:29:53.000Z"},{id:"minimaxai/minimax-m2.1",name:"MiniMax-M2.1",tags:["transformers","safetensors","minimax_m2","text-generation","conversational","custom_code","eval-results","endpoints_compatible"],family:"MiniMax-M2.1",memory:{weights_fp8_gb:229,weights_bf16_gb:458,weights_int4_gb:114.5,kv_per_token_bytes:253952,activation_estimate_gb:761856e-9},version:"2.1",provider:"MiniMax",is_partial:!1,api_pricing:{},architecture:{type:"moe",head_dim:128,num_heads:48,hidden_dim:3072,num_layers:62,vocab_size:200064,num_experts:256,num_kv_heads:8,total_params_b:229,active_params_b:7,intermediate_dim:1536,experts_per_token:8},capabilities:{code:!0,math:!0,tools:!0,vision:!1,reasoning:!0,multilingual:!0,structured_output:!0},release_date:"2025-12-23",context_window:196608,huggingface_id:"MiniMaxAI/MiniMax-M2.1",default_precision:"fp8",supported_frameworks:["vllm","sglang"],supported_precisions:["fp8"],status:"active"},{id:"minimaxai/minimax-m2.5",name:"MiniMax-M2.5",tags:["transformers","safetensors","minimax_m2","text-generation","conversational","custom_code","eval-results","endpoints_compatible"],family:"MiniMax-M2",memory:{weights_fp8_gb:229,weights_bf16_gb:458,weights_int4_gb:114.5,kv_per_token_bytes:253952,activation_estimate_gb:761856e-9},version:"2.5",provider:"MiniMax",is_partial:!1,api_pricing:{minimax:{input_per_m:.3,output_per_m:1.2}},architecture:{type:"moe",head_dim:128,num_heads:48,hidden_dim:3072,num_layers:62,vocab_size:200064,num_experts:256,num_kv_heads:8,total_params_b:229,active_params_b:7,intermediate_dim:1536,experts_per_token:8},capabilities:{code:!0,math:!0,tools:!0,vision:!1,reasoning:!0,multilingual:!0,structured_output:!0},release_date:"2026-02-12",context_window:196608,huggingface_id:"MiniMaxAI/MiniMax-M2.5",default_precision:"fp8",supported_frameworks:["vllm","sglang"],supported_precisions:["fp8"],status:"active"},{id:"minimaxai/minimax-m2.7",name:"MiniMax M2.7",provider:"MiniMax",family:"MiniMax",version:"M2.7",release_date:"2026-04-09",release_date_source:"https://huggingface.co/MiniMaxAI/MiniMax-M2.7",architecture:{type:"moe",total_params_b:228.7,active_params_b:11,num_layers:62,hidden_dim:3072,num_heads:48,num_kv_heads:8,head_dim:128,intermediate_dim:1536,vocab_size:200064,num_experts:256,experts_per_token:8},context_window:204800,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:426,weights_fp8_gb:213,weights_int4_gb:106.5,kv_per_token_bytes:253952,activation_estimate_gb:.001},api_pricing:{openrouter:{input_per_m:.3,output_per_m:1.2}},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang"],tags:["chat","open-weight","moe","code"],huggingface_id:"MiniMaxAI/MiniMax-M2.7",huggingface_id_status:"auto_resolved",license:"other"},{id:"minimaxai/minimax-m2",name:"MiniMax-M2",tags:["transformers","safetensors","minimax_m2","text-generation","conversational","custom_code","eval-results","endpoints_compatible"],family:"MiniMax-M2",memory:{weights_fp8_gb:229,weights_bf16_gb:458,weights_int4_gb:114.5,kv_per_token_bytes:253952,activation_estimate_gb:761856e-9},version:"2",provider:"MiniMax",is_partial:!1,api_pricing:{minimax:{input_per_m:.3,output_per_m:1.1}},architecture:{type:"moe",head_dim:128,num_heads:48,hidden_dim:3072,num_layers:62,vocab_size:200064,num_experts:256,num_kv_heads:8,total_params_b:229,active_params_b:7,intermediate_dim:1536,experts_per_token:8},capabilities:{code:!0,math:!0,tools:!0,vision:!1,reasoning:!0,multilingual:!0,structured_output:!0},release_date:"2025-10-26",context_window:196608,huggingface_id:"MiniMaxAI/MiniMax-M2",default_precision:"fp8",supported_frameworks:["vllm","sglang"],supported_precisions:["fp8"],status:"active"},{id:"minimaxai/minimax-m3",name:"MiniMax M3",provider:"MiniMax",family:"MiniMax",ver

1sion:"M3",release_date:"2026-06-02",release_date_source:"https://huggingface.co/MiniMaxAI/MiniMax-M3",architecture:{type:"moe",total_params_b:427,active_params_b:25.2,num_layers:60,hidden_dim:6144,num_heads:64,num_kv_heads:4,head_dim:128,intermediate_dim:3072,vocab_size:200064,num_experts:128,experts_per_token:4,shared_experts:1},context_window:1048576,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:795.4,weights_fp8_gb:397.7,weights_int4_gb:198.9,kv_per_token_bytes:122880,activation_estimate_gb:.001},api_pricing:{openrouter:{input_per_m:.3,output_per_m:1.2}},capabilities:{tools:!0,vision:!0,structured_output:!1,code:!0,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang"],tags:["chat","open-weight","moe","vision","code"],huggingface_id:"MiniMaxAI/MiniMax-M3",huggingface_id_status:"auto_resolved",license:"other"},{id:"mistral/codestral-22b",name:"Codestral 22B",provider:"Mistral AI",family:"Codestral",version:"1",release_date:"2024-05-29",architecture:{type:"dense",total_params_b:22,active_params_b:22,num_layers:56,hidden_dim:6144,num_heads:48,num_kv_heads:8,head_dim:128,intermediate_dim:16384,vocab_size:32768},context_window:32768,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:44,weights_fp8_gb:22,weights_int4_gb:11,kv_per_token_bytes:229376,activation_estimate_gb:1.5},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!0,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm"],api_pricing:{mistral:{input_per_m:.3,output_per_m:.9},together:{input_per_m:.9,output_per_m:.9}},tags:["code","code-specialized","open-weight"],huggingface_id:"mistralai/Codestral-22B-v0.1",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/mistralai/Codestral-22B-v0.1",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.184Z",license:"other"},{id:"mistral/codestral-mamba-7b",name:"Codestral Mamba 7B",provider:"Mistral AI",family:"Codestral",version:"1",release_date:"2024-07-16",architecture:{type:"hybrid",total_params_b:7.3,active_params_b:7.3,num_layers:64,hidden_dim:4096,num_heads:1,num_kv_heads:1,head_dim:4096,intermediate_dim:8192,vocab_size:32768},context_window:262144,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:14.6,weights_fp8_gb:7.3,weights_int4_gb:3.65,kv_per_token_bytes:0,activation_estimate_gb:1},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!0,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang"],api_pricing:{mistral:{input_per_m:.2,output_per_m:.6}},tags:["code","code-specialized","ssm","mamba","open-weight"],huggingface_id:"mistralai/Mamba-Codestral-7B-v0.1",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/mistralai/Mamba-Codestral-7B-v0.1",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.184Z",license:"apache-2.0"},{id:"mistral/ministral-8b",name:"Ministral 8B",provider:"Mistral AI",family:"Ministral",version:"1",release_date:"2024-10-16",architecture:{type:"dense",total_params_b:8,active_params_b:8,num_layers:36,hidden_dim:4096,num_heads:32,num_kv_heads:8,head_dim:128,intermediate_dim:14336,vocab_size:32768},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:16,weights_fp8_gb:8,weights_int4_gb:4,kv_per_token_bytes:147456,activation_estimate_gb:1},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{mistral:{input_per_m:.1,output_per_m:.1},openrouter:{input_per_m:.15,output_per_m:.15}},tags:["chat","code","open-weight","lightweight"],huggingface_id:"mistralai/Ministral-3-8B-Instruct-2512",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/mistralai/Ministral-3-8B-Instruct-2512",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.184Z",license:"apache-2.0"},{id:"mistral/mistral-7b",name:"Mistral 7B",provider:"Mistral AI",family:"Mistral",version:"0.1",release_date:"2023-09-27",architecture:{type:"dense",total_params_b:7.3,active_params_b:7.3,num_layers:32,hidden_dim:4096,num_heads:32,num_kv_heads:8,head_dim:128,intermediate_dim:14336,vocab_size:32e3},context_window:32768,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:14.6,weights_fp8_gb:7.3,weights_int4_gb:3.65,kv_per_token_bytes:131072,activation_estimate_gb:1},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{together:{input_per_m:.2,output_per_m:.2},deepinfra:{input_per_m:.07,output_per_m:.07},openrouter:{input_per_m:.11,output_per_m:.19}},tags:["chat","code","open-weight"],huggingface_id:"mistralai/Mistral-7B-Instruct-v0.3",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.3",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.184Z",license:"apache-2.0"},{id:"mistral/mistral-large-2411",name:"Mistral Large 2411",provider:"Mistral AI",family:"Mistral Large",version:"2411",release_date:"2024-11-18",architecture:{type:"dense",total_params_b:123,active_params_b:123,num_layers:88,hidden_dim:12288,num_heads:96,num_kv_heads:8,head_dim:128,intermediate_dim:28672,vocab_size:32768},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:246,weights_fp8_gb:123,weights_int4_gb:61.5,kv_per_token_bytes:360448,activation_estimate_gb:3},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm"],api_pricing:{mistral:{input_per_m:2,output_per_m:6},openrouter:{input_per_m:2,output_per_m:6}},tags:["chat","code","reasoning","open-weight","flagship"],huggingface_id:"mistralai/Mistral-Large-Instruct-2411",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/mistralai/Mistral-Large-Instruct-2411",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.184Z",license:"other"},{id:"mistral/mistral-large",name:"Mistral Large 2",provider:"Mistral AI",family:"Mistral Large",version:"2",release_date:"2024-0

17-24",architecture:{type:"dense",total_params_b:123,active_params_b:123,num_layers:88,hidden_dim:12288,num_heads:96,num_kv_heads:8,head_dim:128,intermediate_dim:28672,vocab_size:32768},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:246,weights_fp8_gb:123,weights_int4_gb:61.5,kv_per_token_bytes:360448,activation_estimate_gb:3.5},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm"],api_pricing:{mistral:{input_per_m:2,output_per_m:6},together:{input_per_m:2.5,output_per_m:2.5},openrouter:{input_per_m:2,output_per_m:6}},tags:["chat","code","instruction-following","open-weight","flagship"],huggingface_id:"mistralai/Mistral-Large-Instruct-2407",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/mistralai/Mistral-Large-Instruct-2407",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.184Z",license:"other"},{id:"mistral/mistral-medium-3",name:"Mistral Medium 3",provider:"Mistral AI",family:"Mistral",version:"medium-3",release_date:"2025-05-08",architecture:{type:"dense",total_params_b:70,active_params_b:70,num_layers:80,hidden_dim:8192,num_heads:64,num_kv_heads:8,head_dim:128,intermediate_dim:28672,vocab_size:131072},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:140,weights_fp8_gb:70,weights_int4_gb:35,kv_per_token_bytes:655360,activation_estimate_gb:3},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm"],api_pricing:{mistral:{input_per_m:2,output_per_m:6},openrouter:{input_per_m:.4,output_per_m:2}},tags:["chat","code","reasoning","open-weight","estimated-architecture"],huggingface_id:"mistralai/Mistral-Medium-3.5-128B",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/mistralai/Mistral-Medium-3.5-128B",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.184Z",license:"other"},{id:"mistral/mistral-nemo-12b",name:"Mistral Nemo 12B",provider:"Mistral AI",family:"Mistral Nemo",version:"1",release_date:"2024-07-18",architecture:{type:"dense",total_params_b:12,active_params_b:12,num_layers:40,hidden_dim:5120,num_heads:32,num_kv_heads:8,head_dim:128,intermediate_dim:14336,vocab_size:131072},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:24,weights_fp8_gb:12,weights_int4_gb:6,kv_per_token_bytes:163840,activation_estimate_gb:1.5},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{together:{input_per_m:.18,output_per_m:.18},deepinfra:{input_per_m:.13,output_per_m:.13},openrouter:{input_per_m:.02,output_per_m:.03}},tags:["chat","code","instruction-following","open-weight"],huggingface_id:"nvidia/Mistral-NeMo-12B-Instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/nvidia/Mistral-NeMo-12B-Instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.184Z",license:"apache-2.0"},{id:"mistral/mistral-small-24b",name:"Mistral Small 24B",provider:"Mistral AI",family:"Mistral Small",version:"3",release_date:"2025-01-29",architecture:{type:"dense",total_params_b:24,active_params_b:24,num_layers:40,hidden_dim:6144,num_heads:48,num_kv_heads:8,head_dim:128,intermediate_dim:16384,vocab_size:32768},context_window:32768,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:48,weights_fp8_gb:24,weights_int4_gb:12,kv_per_token_bytes:163840,activation_estimate_gb:1.5},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{mistral:{input_per_m:.1,output_per_m:.3},together:{input_per_m:.3,output_per_m:.3},openrouter:{input_per_m:.075,output_per_m:.2}},tags:["chat","code","instruction-following","open-weight"],huggingface_id:"mistralai/Devstral-Small-2-24B-Instruct-2512",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/mistralai/Devstral-Small-2-24B-Instruct-2512",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.184Z",license:"apache-2.0"},{id:"mistral/mistral-small-3.1-24b",name:"Mistral Small 3.1 24B",provider:"Mistral AI",family:"Mistral Small",version:"3.1",release_date:"2025-03-18",architecture:{type:"dense",total_params_b:24,active_params_b:24,num_layers:40,hidden_dim:5120,num_heads:32,num_kv_heads:8,head_dim:128,intermediate_dim:14336,vocab_size:32768},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:48,weights_fp8_gb:24,weights_int4_gb:12,kv_per_token_bytes:163840,activation_estimate_gb:1.5},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{mistral:{input_per_m:.1,output_per_m:.3},openrouter:{input_per_m:.35,output_per_m:.56}},tags:["chat","code","multimodal","open-weight"],huggingface_id:"mistralai/Mistral-Small-3.1-24B-Instruct-2503",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/mistralai/Mistral-Small-3.1-24B-Instruct-2503",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.184Z",license:"apache-2.0"},{id:"mistral/mixtral-8x22b",name:"Mixtral 8x22B",provider:"Mistral AI",family:"Mixtral",version:"8x22b",release_date:"2024-04-17",architecture:{type:"moe",total_params_b:141,active_params_b:39,num_layers:56,hidden_dim:6144,num_heads:48,num_kv_heads:8,head_dim:128,intermediate_dim:16384,vocab_size:32768,num_experts:8,experts_per_token:2},context_window:65536,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:282,weights_fp8_gb:141,weights_int4_gb:70.5,kv_per_token_bytes:229376,activation_estimate_gb:2.5},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm"],api_pricing:{mistral:{input_per_m:2,output_per_m:6},together:{input_per_m:1.2,output_per_m:1.2},openrouter:{input_per_m:2,output_per_m:6},fireworks:{input_per_m:1.2,output_per_m:1.2}},tags:["chat","code","moe","open-weight"],release_date_source:"https://docs.mistral.ai/getting-started/models/",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-18T13:16:54.712Z"},{id:"mistral/mixtral-8x7b-instruct",name:"Mixtral 8x7B Instruct",provider:"Mistral AI",family:"Mixtral",version:"8x7b-instruct",release_date:"2024-01-08",architecture:{type:"moe",total_params_b:46.7,active_params_b:12.9,num_layers:32,hidden_dim:4096,num_heads:32,num_kv_heads:8,head_dim:128,intermediate_dim:14336,vocab_size:32e3,num_experts:8,experts_per_token:2},context_window:32768,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:93.4,weights_fp8_gb:46.7,weights_int4_gb:23.35,kv_per_token_bytes:131072,activation_estimate_gb:1.5},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{together:{input_per_m:.24,output_per_m:.24},groq:{input_per_m:.24,output_per_m:.24}},tags:["chat","code","moe","instruct","open-weight"],huggingface_id:"mistralai/Mixtral-8x7B-Instruct-v0.1",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/mistralai/Mixtral-8x7B-Instruct-v0.1",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.184Z",license:"apache-2.0"},{id:"mistral/mixtral-8x7b",name:"Mixtral 8x7B",provider:"Mistral AI",family:"Mixtral",version:"8x7b",release_date:"2023-12-11",architecture:{type:"moe",total_params_b:46.7,active_params_b:12.9,num_layers:32,hidden_dim:4096,num_heads:32,num_kv_heads:8,head_dim:128,intermediate_dim:14336,vocab_size:32e3,num_experts:8,experts_per_token:2},context_window:32768,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:93.4,weights_fp8_gb:46.7,weights_int4_gb:23.35,kv_per_token_bytes:131072,activation_estimate_gb:1.5},capabilities:{tools:!1,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{together:{input_per_m:.6,output_per_m:.6},fireworks:{input_per_m:.5,output_per_m:.5}},tags:["chat","code","moe","open-weight"],huggingface_id:"mistralai/Mixtral-8x7B-Instruct-v0.1",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/mistralai/Mixtral-8x7B-Instruct-v0.1",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.185Z",license:"apache-2.0"},{id:"mistral/pixtral-12b",name:"Pixtral 12B",provider:"Mistral AI",family:"Pixtral",version:"1",release_date:"2024-09-17",architecture:{type:"dense",total_params_b:12,active_params_b:12,num_layers:40,hidden_dim:4096,num_heads:32,num_kv_heads:8,head_dim:128,intermediate_dim:14336,vocab_size:32768},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:24,weights_fp8_gb:12,weights_int4_gb:6,kv_per_token_bytes:163840,activation_estimate_gb:1},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm"],api_pricing:{mistral:{input_per_m:.15,output_per_m:.15}},tags:["chat","vision","vision-language","open-weight"],huggingface_id:"mistralai/Pixtral-12B-2409",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/mistralai/Pixtral-12B-2409",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.185Z",license:"apache-2.0"},{id:"mistralai/mistral-large-3-675b",name:"Mistral Large 3",provider:"Mistral AI",family:"Mistral",version:"Large 3",is_partial:!0,release_date:"2025-12-01",release_date_source:"https://huggingface.co/mistralai/Mistral-Large-3-675B-Instruct-2512",release_date_verified_at:"2026-08-10T20:24:32.980Z",last_verified_at:"2026-08-10T20:24:32.980Z",huggingface_id:"mistralai/Mistral-Large-3-675B-Instruct-2512",license:"apache-2.0",capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},deployment:"api",context_window:262144,api_pricing:{mistral:{input_per_m:.5,output_per_m:1.5}},description:"Mistral Large 3 is Mistral AI's flagship model (mistral-large-2512) — a large, tool-using, reasoning-capable model with a 256K-token context window, served through the Mistral (La Plateforme) API at $0.5/M input, $1.5/M output."},{id:"mistralai/mistral-medium-3.5-128b",name:"Mistral Medium 3.5",provider:"Mistral AI",family:"Mistral",version:"Medium 3.5",release_date:"2026-03-31",release_date_source:"https://huggingface.co/mistralai/Mistral-Medium-3.5-128B",architecture:{type:"dense",total_params_b:127.7,active_params_b:127.7,num_layers:88,hidden_dim:12288,num_heads:96,num_kv_heads:8,head_dim:128,intermediate_dim:28672,vocab_size:131072},context_window:262144,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:237.9,weights_fp8_gb:118.9,weights_int4_gb:59.5,kv_per_token_bytes:360448,activation_estimate_gb:.004},api_pricing:{openrouter:{input_per_m:1.5,output_per_m:7.5}},capabilities:{tools:!1,vision:!0,structured_output:!1,code:!1,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang"],tags:["chat","open-weight","vision"],huggingface_id:"mistralai/Mistral-Medium-3.5-128B",huggingface_id_status:"auto_resolved",license:"other"}

1,{id:"mistralai/mistral-small-4-119b-2603",name:"Mistral Small 4",provider:"Mistral AI",family:"Mistral",version:"Small 4",release_date:"2026-01-23",release_date_source:"https://huggingface.co/mistralai/Mistral-Small-4-119B-2603",architecture:{type:"moe",total_params_b:119.4,active_params_b:6.6,num_layers:36,hidden_dim:4096,num_heads:32,num_kv_heads:32,head_dim:128,intermediate_dim:12288,vocab_size:131072,num_experts:128,experts_per_token:4,shared_experts:1},context_window:1048576,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:222.4,weights_fp8_gb:111.2,weights_int4_gb:55.6,kv_per_token_bytes:589824,activation_estimate_gb:.001},api_pricing:{openrouter:{input_per_m:.15,output_per_m:.6}},capabilities:{tools:!1,vision:!0,structured_output:!1,code:!1,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang"],tags:["chat","open-weight","moe","vision"],huggingface_id:"mistralai/Mistral-Small-4-119B-2603",huggingface_id_status:"auto_resolved",license:"apache-2.0"},{id:"moonshot/kimi-k2.5",name:"Kimi K2.5",provider:"Moonshot AI",family:"Kimi",version:"K2.5",release_date:"2026-01-27",release_date_source:"https://www.kimi.com/ai-models/kimi-k2-5",release_date_verified_at:"2026-05-13T14:42:10Z",last_verified_at:"2026-06-25T15:18:14.185Z",architecture:{type:"moe",total_params_b:1e3,active_params_b:32,num_layers:64,hidden_dim:6144,num_heads:48,num_kv_heads:8,head_dim:128,intermediate_dim:16384,vocab_size:131072,num_experts:256,experts_per_token:8,shared_experts:1},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"fp8",memory:{weights_bf16_gb:2e3,weights_fp8_gb:1e3,weights_int4_gb:500,kv_per_token_bytes:262144,activation_estimate_gb:3},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm","sglang"],api_pricing:{moonshot:{input_per_m:.6,output_per_m:2.4},novita:{input_per_m:.6,output_per_m:3},fireworks:{input_per_m:0,output_per_m:0},featherless:{input_per_m:0,output_per_m:0},openrouter:{input_per_m:.44,output_per_m:2}},tags:["chat","code","reasoning","moe","open-weight","flagship"],huggingface_id:"moonshotai/Kimi-K2.5",huggingface_id_status:"auto_resolved",license:"other"},{id:"moonshotai/kimi-k2-thinking",name:"Kimi-K2-Thinking",tags:["transformers","safetensors","kimi_k2","text-generation","conversational","custom_code","eval-results","endpoints_compatible"],family:"moonshotai",memory:{weights_fp8_gb:1058,weights_bf16_gb:2116,weights_int4_gb:529,kv_per_token_bytes:1748992,activation_estimate_gb:.001748992},version:"2",provider:"moonshotai",api_pricing:{},architecture:{type:"moe",head_dim:112,num_heads:64,hidden_dim:7168,num_layers:61,vocab_size:163840,num_experts:384,num_kv_heads:64,total_params_b:1058,active_params_b:7,intermediate_dim:2048,experts_per_token:8},capabilities:{code:!0,math:!1,tools:!1,vision:!1,reasoning:!1,multilingual:!1,structured_output:!1},release_date:"2025-11-04",context_window:262144,huggingface_id:"moonshotai/Kimi-K2-Thinking",default_precision:"bf16",supported_frameworks:["vllm"],supported_precisions:["bf16"],status:"draft",license:"other",last_verified_at:"2026-06-25T15:18:14.185Z"},{id:"moonshotai/kimi-k2.6",name:"Kimi-K2.6",family:"moonshotai",license:"other",version:"2.6",provider:"moonshotai",is_partial:!0,capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},release_date:"2026-04-14",huggingface_id:"moonshotai/Kimi-K2.6",last_verified_at:"2026-06-23T11:20:52.135Z",release_date_source:"https://huggingface.co/moonshotai/Kimi-K2.6",release_date_verified_at:"2026-06-23T11:20:52.135Z",deployment:"api",api_pricing:{moonshot:{input_per_m:.95,output_per_m:4}},description:"Kimi K2.6 is Moonshot AI's multimodal model built for long-horizon coding and multi-agent orchestration, with a 256K-token context window, served through the Moonshot (Kimi) API.",context_window:262144},{id:"moonshotai/kimi-k2.7-code",name:"Kimi K2.7 Code",provider:"Moonshot AI",family:"Kimi",version:"K2.7",release_date:"2026-06-11",release_date_source:"https://huggingface.co/moonshotai/Kimi-K2.7-Code",architecture:{type:"moe",total_params_b:1026.9,active_params_b:32.9,num_layers:61,hidden_dim:7168,num_heads:64,num_kv_heads:64,head_dim:112,intermediate_dim:18432,vocab_size:163840,num_experts:384,experts_per_token:8,shared_experts:1},context_window:262144,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:1912.7,weights_fp8_gb:956.4,weights_int4_gb:478.2,kv_per_token_bytes:1748992,activation_estimate_gb:.002},api_pricing:{openrouter:{input_per_m:.7,output_per_m:3.5}},capabilities:{tools:!1,vision:!0,structured_output:!1,code:!0,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang"],tags:["chat","open-weight","moe","vision","code"],huggingface_id:"moonshotai/Kimi-K2.7-Code",huggingface_id_status:"auto_resolved",license:"other"},{id:"moonshotai/kimi-k3",name:"Kimi K3",provider:"Moonshot AI",family:"Kimi",version:"K3",release_date:"2026-07-16",release_date_source:"https://huggingface.co/moonshotai/Kimi-K3",architecture:{type:"moe",total_params_b:2779.9,active_params_b:32,num_layers:93,hidden_dim:7168,num_heads:96,num_kv_heads:96,head_dim:75,intermediate_dim:33792,vocab_size:163840,num_experts:896,experts_per_token:16,shared_experts:2},context_window:1048576,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:5178,weights_fp8_gb:2589,weights_int4_gb:1294.5,kv_per_token_bytes:2678400,activation_estimate_gb:.003},api_pricing:{openrouter:{input_per_m:3,output_per_m:15}},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm","sglang"],tags:["chat","code","reasoning","moe","vision","open-weight","flagship"],huggingface_id:"moonshotai/Kimi-K3",huggingface_id_status:"auto_resolved",license:"other"},{id:"moonshotai/kimi-linear-48b-a3b-base",name:"Kimi-Linear-48B-A3B-Base",family:"moonshotai",license:"mit",provider:"moonshotai",capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!0,multilingual:!0,reasoning:!1},release_date:"2025-10-30",release_date_source:"https://huggingface.co/moonshotai/Kimi-Linear-48B-A3B-Base",huggingface_id:"moonshotai/Kimi-Linear-48B-A3B-Base",release_date_verified_at:"2026-06-23T11:28:28.861Z",last_verified_at:"2026-06-23T11:28:28.861Z",version:"48B-A3B-Base",context_window:1048576,architecture:{type:"moe",total_params_b:48,active_params_b:3,num_layers:27,hidden_dim:2304,num_heads:32,num_kv_heads:32,head_dim:72,intermediate_dim:9216,vocab_size:163840,num_experts:256,experts_per_token:8},memory:{weights_bf16_gb:96,weights_fp8_gb:48,weights_int4_gb:24,kv_per_token_bytes:64512,activation_estimate_gb:6},supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",supp

1orted_frameworks:["vllm","sglang"],api_pricing:{},description:"Kimi Linear 48B A3B (Base) is Moonshot AI's open-weight hybrid linear-attention MoE base model — 48B total / 3B active params, 256 experts (8 active), 27 layers (20 linear KDA + 7 full-attention), 1M-token context. Base (pre-instruct) checkpoint for self-hosting.",tags:["moe","linear-attention","base-model","open-weight","long-context"]},{id:"mosaicml/mpt-30b",name:"MPT 30B",provider:"MosaicML",family:"MPT",version:"30b",release_date:"2023-06-22",architecture:{type:"dense",total_params_b:30,active_params_b:30,num_layers:48,hidden_dim:7168,num_heads:64,num_kv_heads:8,head_dim:112,intermediate_dim:19456,vocab_size:50432},context_window:8192,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:60,weights_fp8_gb:30,weights_int4_gb:15,kv_per_token_bytes:344064,activation_estimate_gb:1.5},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","tgi"],api_pricing:{},tags:["chat","code","open-weight"],huggingface_id:"TheBloke/mpt-30B-instruct-GGML",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/TheBloke/mpt-30B-instruct-GGML",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.185Z",license:"cc-by-sa-3.0"},{id:"mosaicml/mpt-7b",name:"MPT 7B",provider:"MosaicML",family:"MPT",version:"7b",release_date:"2023-05-05",architecture:{type:"dense",total_params_b:6.7,active_params_b:6.7,num_layers:32,hidden_dim:4096,num_heads:32,num_kv_heads:32,head_dim:128,intermediate_dim:11008,vocab_size:50432},context_window:65536,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:13.4,weights_fp8_gb:6.7,weights_int4_gb:3.35,kv_per_token_bytes:262144,activation_estimate_gb:.8},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","tgi","ollama"],api_pricing:{},tags:["chat","code","open-weight"],huggingface_id:"gl198976/mpt-7b-instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/gl198976/mpt-7b-instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.185Z",license:"cc-by-sa-3.0"},{id:"nomic/nomic-embed-text-v1.5",name:"Nomic Embed Text v1.5",provider:"Nomic AI",family:"Nomic Embed",version:"1.5",release_date:"2024-02-07",architecture:{type:"dense",total_params_b:.137,active_params_b:.137,num_layers:12,hidden_dim:768,num_heads:12,num_kv_heads:12,head_dim:64,intermediate_dim:3072,vocab_size:30528},context_window:8192,supported_precisions:["bf16","fp8"],default_precision:"bf16",memory:{weights_bf16_gb:.274,weights_fp8_gb:.137,weights_int4_gb:.069,kv_per_token_bytes:36864,activation_estimate_gb:.05},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","tgi","ollama"],api_pricing:{nomic:{input_per_m:.008,output_per_m:.008}},tags:["embedding","lightweight","open-weight"],huggingface_id:"nomic-ai/nomic-embed-text-v1.5",huggingface_id_status:"manual",release_date_source:"https://huggingface.co/nomic-ai/nomic-embed-text-v1.5",last_verified_at:"2026-06-25T16:19:04.534Z",license:"apache-2.0"},{id:"nousresearch/hermes-3-70b",name:"Hermes 3 70B",provider:"Nous Research",fam

1ily:"Hermes 3",version:"3",release_date:"2024-08-19",architecture:{type:"dense",total_params_b:70.6,active_params_b:70.6,num_layers:80,hidden_dim:8192,num_heads:64,num_kv_heads:8,head_dim:128,intermediate_dim:28672,vocab_size:128256},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:141.2,weights_fp8_gb:70.6,weights_int4_gb:35.3,kv_per_token_bytes:327680,activation_estimate_gb:2.5},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm"],api_pricing:{together:{input_per_m:.88,output_per_m:.88}},tags:["chat","code","instruction-following","open-weight"],huggingface_id:"NousResearch/Hermes-2-Theta-Llama-3-70B",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/NousResearch/Hermes-2-Theta-Llama-3-70B",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.185Z",license:"llama3"},{id:"nousresearch/hermes-3-8b",name:"Hermes 3 8B",provider:"Nous Research",fam

1ily:"Hermes 3",version:"3",release_date:"2024-08-19",architecture:{type:"dense",total_params_b:8.03,active_params_b:8.03,num_layers:32,hidden_dim:4096,num_heads:32,num_kv_heads:8,head_dim:128,intermediate_dim:14336,vocab_size:128256},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:16.06,weights_fp8_gb:8.03,weights_int4_gb:4.015,kv_per_token_bytes:131072,activation_estimate_gb:1},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{together:{input_per_m:.18,output_per_m:.18}},tags:["chat","code","instruction-following","open-weight"],huggingface_id:"NousResearch/Hermes-3-Llama-3.1-8B",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/NousResearch/Hermes-3-Llama-3.1-8B",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.185Z",license:"llama3"},{id:"nvidia/alpamayo-10b",name:"Alpamayo 1.5-10B",provider:"NVIDIA",family:"Alpamayo",version:"1.5",release_date:"2025-04-01",architecture:{type:"dense",total_params_b:10,active_params_b:10,num_layers:32,hidden_dim:4096,num_heads:32,num_kv_heads:8,head_dim:128,intermediate_dim:11008,vocab_size:128256},context_window:8192,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:20,weights_fp8_gb:10,weights_int4_gb:5,kv_per_token_bytes:131072,activation_estimate_gb:1},capabilities:{tools:!1,vision:!0,structured_output:!1,code:!1,math:!1,multilingual:!1,reasoning:!0},supported_frameworks:["vllm"],api_pricing:{},tags:["autonomous-driving","vision","trajectory-prediction","open-weight"],huggingface_id:"nvidia/Alpamayo-1.5-10B",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/nvidia/Alpamayo-1.5-10B",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.185Z",license:"other"},{id:"nvidia/bionemo-esm2-650m",name:"BioNeMo ESM-2 650M",provider:"NVIDIA",family:"BioNeMo",version:"1.0",modality:"protein",release_date:"2023-11-13",release_date_source:"https://docs.nvidia.com/bionemo-framework/latest/models/esm2-nv.html",release_date_verified_at:"2026-06-25T16:00:00Z",last_verified_at:"2026-06-25T16:00:00Z",huggingface_id:"facebook/esm2_t33_650M_UR50D",huggingface_id_status:"manual",license:"mit",specialized_architecture:{type:"encoder",total_params_b:.65,num_layers:33,hidden_dim:1280,notes:"Protein language model — masked-language transformer over amino-acid sequences (UniRef50 corpus). Vocabulary is 33 tokens (20 canonical amino acids + special tokens). Produces per-residue contextual embeddings used for structure prediction, variant-effect prediction, and embedding-based search. Originally published by Meta (ESM-2); BioNeMo packages it with NVIDIA-optimized training + inference for drug-discovery workflows."},specialized_memory:{weights_bf16_gb:1.3,weights_fp8_gb:.65,activation_estimate_gb:.8},specialized_pricing:{kind:"free_open_weights"},description:"ESM-2 650M is a 33-layer transformer trained as a protein language model on ~65M UniRef50 sequences. NVIDIA BioNeMo provides containerized inference (NIM) and training-optimized FlashAttention kernels for the model.",intended_use:"Per-residue protein embeddings for structure prediction (downstream models like ESMFold), variant-effect prediction, evolutionary search, drug-target identification. Self-host via NeMo or NIM container.",tags:["protein","biology","drug-discovery","embeddings","open-weight","esm"]},{id:"nvidia/canary-1b",name:"Canary 1B",provider:"NVIDIA",family:"Canary",version:"1",release_date:"2024-04-18",architecture:{type:"dense",total_params_b:1,active_params_b:1,num_layers:24,hidden_dim:1024,num_heads:8,num_kv_heads:8,head_dim:128,intermediate_dim:4096,vocab_size:128e3},context_window:4096,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:2,weights_fp8_gb:1,weights_int4_gb:.5,kv_per_token_bytes:12288,activation_estimate_gb:.2},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["tensorrt-llm","vllm"],api_pricing:{"nvidia-nim":{input_per_m:.04,output_per_m:.04}},tags:["nvidia","open-source","asr","speech-recognition","multilingual"],huggingface_id:"nvidia/canary-1b-v2",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/nvidia/canary-1b-v2",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.185Z",license:"cc-by-4.0"},{id:"nvidia/cosmos-7b",name:"Cosmos 7B",provider:"NVIDIA",family:"Cosmos",version:"1",release_date:"2025-01-07",architecture:{type:"dense",total_params_b:7,active_params_b:7,num_layers:28,hidden_dim:3072,num_heads:24,num_kv_heads:8,head_dim:128,intermediate_dim:8192,vocab_size:64e3},context_window:4096,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:14,weights_fp8_gb:7,weights_int4_gb:3.5,kv_per_token_bytes:114688,activation_estimate_gb:.8},capabilities:{tools:!1,vision:!0,structured_output:!1,code:!1,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["tensorrt-llm"],api_pricing:{},tags:["world-model","video-generation","physical-ai","open-weight"],huggingface_id:"nvidia/Cosmos-1.0-Diffusion-7B-Video2World",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/nvidia/Cosmos-1.0-Diffusion-7B-Video2World",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.185Z",license:"other"},{id:"nvidia/eagle-2-1b",name:"Eagle 2 1B",provider:"NVIDIA",family:"Eagle",version:"2",release_date:"2025-01-15",architecture:{type:"dense",total_params_b:1.3,active_params_b:1.3,num_layers:24,hidden_dim:2048,num_heads:16,num_kv_heads:16,head_dim:128,intermediate_dim:5632,vocab_size:151936},context_window:4096,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:2.6,weights_fp8_gb:1.3,weights_int4_gb:.65,kv_per_token_bytes:196608,activation_estimate_gb:.2},capabilities:{tools:!1,vision:!0,structured_output:!1,code:!1,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","ollama"],api_pricing:{},tags:["vision","edge","open-weight"],huggingface_id:"nvidia/Eagle2-1B",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/nvidia/Eagle2-1B",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.185Z",license:"cc-by-nc-4.0"}

1,{id:"nvidia/eagle-2-9b",name:"Eagle 2 9B",provider:"NVIDIA",family:"Eagle",version:"2",release_date:"2025-01-15",architecture:{type:"dense",total_params_b:9,active_params_b:9,num_layers:32,hidden_dim:4096,num_heads:32,num_kv_heads:8,head_dim:128,intermediate_dim:14336,vocab_size:151936},context_window:8192,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:18,weights_fp8_gb:9,weights_int4_gb:4.5,kv_per_token_bytes:131072,activation_estimate_gb:.5},capabilities:{tools:!1,vision:!0,structured_output:!1,code:!1,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","tensorrt-llm"],api_pricing:{},tags:["vision","multimodal","open-weight"],huggingface_id:"nvidia/Eagle2-9B",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/nvidia/Eagle2-9B",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.185Z",license:"cc-by-nc-4.0"},{id:"nvidia/eagle-2.5-8b",name:"Eagle 2.5 8B",provider:"NVIDIA",family:"Eagle",version:"2.5",release_date:"2025-03-25",architecture:{type:"dense",total_params_b:8,active_params_b:8,num_layers:32,hidden_dim:4096,num_heads:32,num_kv_heads:8,head_dim:128,intermediate_dim:14336,vocab_size:152064},context_window:16384,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:16,weights_fp8_gb:8,weights_int4_gb:4,kv_per_token_bytes:131072,activation_estimate_gb:.5},capabilities:{tools:!1,vision:!0,structured_output:!1,code:!0,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","tensorrt-llm"],api_pricing:{},tags:["vision","multimodal","open-weight","edge"],huggingface_id:"nvidia/Eagle2.5-8B",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/nvidia/Eagle2.5-8B",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.185Z",license:"other"},{id:"nvidia/edify-image",name:"Edify Image",provider:"NVIDIA",family:"Picasso",version:"1.0",modality:"image-gen",release_date:"2024-03-18",release_date_source:"https://developer.nvidia.com/edify",release_date_verified_at:"2026-06-25T16:00:00Z",last_verified_at:"2026-06-25T16:00:00Z",license:"nvidia-picasso-commercial-license",specialized_architecture:{type:"diffusion",total_params_b:4,notes:"Cascaded latent diffusion text-to-image model — T5-style text encoder + base 1024×1024 latent diffusion U-Net + upsampling diffusion stages. Trained on NVIDIA's curated licensed-content corpus (Getty Images, Shutterstock partner data). Available exclusively as a hosted API via NVIDIA Picasso, with optional fine-tuning on customer-licensed imagery. Inference on H100 / L40S with TensorRT optimization."},specialized_memory:{weights_bf16_gb:8,weights_fp8_gb:4,activation_estimate_gb:4},specialized_pricing:{kind:"per_image",price:.04,resolution:"1024x1024"},description:"Edify Image is NVIDIA's enterprise text-to-image model — trained exclusively on commercially-licensed data through the Picasso platform. Positioned for advertising, design, and content-creation workflows where copyright-clean image generation is a hard requirement.",intended_use:"Commercial image generation for marketing, advertising, product photography, and creative tooling where IP-cleared training data is required. Fine-tunable on customer corpora.",tags:["image-gen","diffusion","text-to-image","commercial","picasso","licensed-data"]},{id:"nvidia/helpsteer2-llama-3.1-70b",name:"HelpSteer2 Llama 3.1 70B",provider:"NVIDIA",family:"Llama 3.1",version:"helpsteer2",release_date:"2024-08-20",architecture:{type:"dense",total_params_b:70.6,active_params_b:70.6,num_layers:80,hidden_dim:8192,num_heads:64,num_kv_heads:8,head_dim:128,intermediate_dim:28672,vocab_size:128256},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:141.2,weights_fp8_gb:70.6,weights_int4_gb:35.3,kv_per_token_bytes:327680,activation_estimate_gb:2.5},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["tensorrt-llm","vllm","sglang"],api_pricing:{"nvidia-nim":{input_per_m:.5,output_per_m:.5},openrouter:{input_per_m:.4,output_per_m:.4},together:{input_per_m:.5,output_per_m:.5},fireworks:{input_per_m:.9,output_per_m:.9},groq:{input_per_m:.59,output_per_m:.79}},tags:["nvidia","open-source","alignment","steerlm","helpsteer"],huggingface_id:"nvidia/Llama-3.1-Nemotron-70B-Instruct-HF",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/nvidia/Llama-3.1-Nemotron-70B-Instruct-HF",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.185Z",license:"llama3.1"},{id:"nvidia/llama-3.1-nemotron-51b",name:"Llama 3.1 Nemotron 51B",provider:"NVIDIA",family:"Llama 3.1",version:"3.1-nemotron-51b",release_date:"2024-10-15",architecture:{type:"dense",total_params_b:51,active_params_b:51,num_layers:64,hidden_dim:8192,num_heads:64,num_kv_heads:8,head_dim:128,intermediate_dim:28672,vocab_size:128256},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:102,weights_fp8_gb:51,weights_int4_gb:25.5,kv_per_token_bytes:262144,activation_estimate_gb:2},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["tensorrt-llm","vllm","sglang"],api_pricing:{"nvidia-nim":{input_per_m:.4,output_per_m:.4}},tags:["nvidia","open-source","pruned","minitron","nvidia-nim-optimized"],huggingface_id:"nvidia/Llama-3_1-Nemotron-51B-Instruct",huggingface_id_status:"manual",release_date_source:"https://huggingface.co/nvidia/Llama-3_1-Nemotron-51B-Instruct",last_verified_at:"2026-06-25T16:19:04.534Z",license:"other"},{id:"nvidia/llama-3.1-nemotron-70b-instruct",name:"Llama 3.1 Nemotron 70B Instruct",provider:"NVIDIA",family:"Llama 3.1",version:"3.1-nemotron",release_date:"2024-10-15",architecture:{type:"dense",total_params_b:70.6,active_params_b:70.6,num_layers:80,hidden_dim:8192,num_heads:64,num_kv_heads:8,head_dim:128,intermediate_dim:28672,vocab_size:128256},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:14

11.2,weights_fp8_gb:70.6,weights_int4_gb:35.3,kv_per_token_bytes:327680,activation_estimate_gb:2.5},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm"],api_pricing:{nvidia:{input_per_m:1,output_per_m:1},together:{input_per_m:.88,output_per_m:.88}},tags:["chat","code","instruction-following","rlhf","open-weight"],huggingface_id:"nvidia/Llama-3.1-Nemotron-70B-Instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/nvidia/Llama-3.1-Nemotron-70B-Instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.185Z",license:"llama3.1"},{id:"nvidia/llama-3.1-nemotron-70b-reward",name:"Llama 3.1 Nemotron 70B Reward",provider:"NVIDIA",family:"Llama 3.1",version:"3.1-nemotron-reward",release_date:"2024-10-15",architecture:{type:"dense",total_params_b:70.6,active_params_b:70.6,num_layers:80,hidden_dim:8192,num_heads:64,num_kv_heads:8,head_dim:128,intermediate_dim:28672,vocab_size:128256},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:141.2,weights_fp8_gb:70.6,weights_int4_gb:35.3,kv_per_token_bytes:327680,activation_estimate_gb:2.5},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["tensorrt-llm","vllm","sglang"],api_pricing:{"nvidia-nim":{input_per_m:.5,output_per_m:.5}},tags:["nvidia","open-source","reward-model","rlhf","alignment"],huggingface_id:"nvidia/Llama-3.1-Nemotron-70B-Reward",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/nvidia/Llama-3.1-Nemotron-70B-Reward",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.185Z",license:"llama3.1"},{id:"nvidia/maxine-eye-contact",name:"Maxine Eye Contact",provider:"NVIDIA",family:"Maxine",version:"1.0",modality:"video-fx",release_date:"2023-03-21",release_date_source:"https://developer.nvidia.com/maxine",release_date_verified_at:"2026-06-25T16:00:00Z",last_verified_at:"2026-06-25T16:00:00Z",license:"nvidia-maxine-sdk-license",specialized_architecture:{type:"pipeline",total_params_b:.05,notes:"Real-time video-effects pipeline. Detects the speaker's face + eye region, then warps the eye region per frame so the eyes appear to look directly into the camera even when the user is reading off-screen. Implementation is a cascade of small CNNs — face detector → landmark predictor → eye-region warp generator — packaged as the Maxine Eye Contact effect. Designed for sub-frame latency on RTX 40-series and L40S."},specialized_memory:{weights_bf16_gb:.1,activation_estimate_gb:.05},specialized_pricing:{kind:"per_video_minute",price:0},description:"Maxine Eye Contact is one effect in NVIDIA's Maxine SDK for real-time video communications. Maintains eye-contact appearance during virtual meetings, presentations, and video creation by synthetically redirecting the speaker's gaze toward the camera.",intended_use:"Real-time video conferencing, streaming, video creation tools. Runs on the client GPU (or server-side via Maxine Cloud) at video frame rate.",tags:["video-fx","video-conferencing","real-time","gaze-redirection","sdk"]},{id:"nvidia/megatron-turing-530b",name:"Megatron-Turing NLG 530B",provider:"NVIDIA",family:"Megatron-Turing",version:"1",release_date:"2022-01-28",architecture:{type:"dense",total_params_b:530,active_params_b:530,num_layers:105,hidden_dim:20480,num_heads:128,num_kv_heads:128,head_dim:160,intermediate_dim:53248,vocab_size:50257},context_window:2048,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:1060,weights_fp8_gb:530,weights_int4_gb:265,kv_per_token_bytes:3440640,activation_estimate_gb:12},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["tensorrt-llm","vllm"],api_pricing:{},tags:["nvidia","historical","research","dense","largest"],release_date_source:"https://developer.nvidia.com/blog/using-deepspeed-and-megatron-to-train-megatron-turing-nlg-530b-the-worlds-largest-and-most-powerful-generative-language-model/",license:"nvidia-research-lice

1nse",release_date_verified_at:"2026-06-25T16:20:15.668Z",last_verified_at:"2026-06-25T16:20:15.668Z"},{id:"nvidia/minitron-4b",name:"Minitron 4B",provider:"NVIDIA",family:"Nemotron",version:"1",release_date:"2024-07-16",architecture:{type:"dense",total_params_b:4,active_params_b:4,num_layers:32,hidden_dim:3072,num_heads:24,num_kv_heads:8,head_dim:128,intermediate_dim:9216,vocab_size:256e3},context_window:8192,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:8,weights_fp8_gb:4,weights_int4_gb:2,kv_per_token_bytes:49152,activation_estimate_gb:.5},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["tensorrt-llm","vllm","sglang"],api_pricing:{"nvidia-nim":{input_per_m:.06,output_per_m:.06}},tags:["nvidia","open-source","pruned","distilled","nvidia-nim-optimized"],huggingface_id:"nvidia/Minitron-4B-Base",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/nvidia/Minitron-4B-Base",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.185Z",license:"other"},{id:"nvidia/minitron-8b",name:"Minitron 8B",provider:"NVIDIA",family:"Nemotron",version:"1",release_date:"2024-07-16",architecture:{type:"dense",total_params_b:8,active_params_b:8,num_layers:32,hidden_dim:4096,num_heads:32,num_kv_heads:8,head_dim:128,intermediate_dim:16384,vocab_size:256e3},context_window:8192,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:16,weights_fp8_gb:8,weights_int4_gb:4,kv_per_token_bytes:65536,activation_estimate_gb:.8},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["tensorrt-llm","vllm","sglang"],api_pricing:{"nvidia-nim":{input_per_m:.1,output_per_m:.1}},tags:["nvidia","open-source","pruned","distilled","nvidia-nim-optimized"],huggingface_id:"nvidia/Minitron-8B-Base",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/nvidia/Minitron-8B-Base",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.185Z",license:"other"},{id:"nvidia/nemotron-15b",name:"Nemotron 15B",provider:"NVIDIA",family:"Nemotron",version:"1",release_date:"2024-06-14",architecture:{type:"dense",total_params_b:15,active_params_b:15,num_layers:32,hidden_dim:6144,num_heads:48,num_kv_heads:8,head_dim:128,intermediate_dim:24576,vocab_size:256e3},context_window:4096,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:30,weights_fp8_gb:15,weights_int4_gb:7.5,kv_per_token_bytes:131072,activation_estimate_gb:1.5},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tensorrt-llm"],api_pricing:{nvidia:{input_per_m:.3,output_per_m:.3}},tags:["chat","open-weight"],release_date_source:"https://developer.nvidia.com/blog/nvidia-ai-foundation-models-build-custom-enterprise-chatbots-and-co-pilots-with-production-ready-llms/",license:"nvidia-open-model-license",release_date_verified_at:"2026-06-25T16:20:15.668Z",last_verified_at:"2026-06-25T16:20:15.668Z"},{id:"nvidia/nemotron-3-embed-1b-bf16",name:"Nemotron-3-Embed-1B-BF16",tags:["sentence-transformers","safetensors","ministral3","feature-extraction","text","text-embeddings","retrieval","semantic-search"],family:"nvidia",memory:{weights_fp8_gb:1.1,weights_bf16_gb:2.2,weights_int4_gb:.55,kv_per_token_bytes:65536,activation_estimate_gb:131072e-9},version:"3",provider:"nvidia",is_partial:!1,api_pricing:{},architecture:{type:"dense",head_dim:128,num_heads:24,hidden_dim:2048,num_layers:16,vocab_size:131072,num_kv_heads:8,total_params_b:1.1,active_params_b:1.1,intermediate_dim:6144},capabilities:{code:!1,math:!1,tools:!1,vision:!1,reasoning:!1,multilingual:!1,structured_output:!1},release_date:"2026-01-01",context_window:262144,huggingface_id:"nvidia/Nemotron-3-Embed-1B-BF16",default_precision:"bf16",supported_frameworks:["vllm"],supported_precisions:["bf16","fp8","int4"],status:"active"},{id:"nvidia/nemotron-3-embed-1b-nvfp4",name:"Nemotron-3-Embed-1B-NVFP4",tags:["vllm","safetensors","ministral3","text","text-embeddings","retrieval","semantic-search","nvfp4"],family:"nvidia",version:"3",modality:"image-gen",provider:"nvidia",is_partial:!1,api_pricing:{},description:"wer",capabilities:{code:!1,math:!1,tools:!1,vision:!1,reasoning:!1,multilingual:!1,structured_output:!1},release_date:"2026-01-01",huggingface_id:"nvidia/Nemotron-3-Embed-1B-NVFP4",default_precision:"bf16",specialized_memory:{weights_bf16_gb:2,activation_estimate_gb:0},specialized_pricing:{kind:"free_open_weights"},supported_frameworks:["vllm"],supported_precisions:["bf16","fp8","int4"],specialized_architecture:{type:"encoder",notes:"wer",total_params_b:1},status:"active"},{id:"nvidia/nemotron-3-embed-8b-bf16",name:"Nemotron-3-Embed-8B-BF16",tags:["sentence-transformers","safetensors","ministral3","feature-extraction","text","text-embeddings","retrieval","semantic-search"],family:"nvidia",version:"3",modality:"image-gen",provider:"nvidia",is_partial:!1,api_pricing:{},description:"wer",capabilities:{code:!0,math:!1,tools:!0,vision:!1,reasoning:!1,multilingual:!1,structured_output:!0},release_date:"2026-01-01",huggingface_id:"nvidia/Nemotron-3-Embed-8B-BF16",default_precision:"bf16",specialized_memory:{weights_bf16_gb:2,activation_estimate_gb:0},specialized_pricing:{kind:"free_open_weights"},supported_frameworks:["vllm"],supported_precisions:["bf16","fp8","int4"],specialized_architecture:{type:"encoder",notes:"wkjher",hidden_dim:2,num_layers:2,total_params_b:1,active_params_b:2},status:"active"},{id:"nvidia/nemotron-340b",name:"Nemotron 340B",provider:"NVIDIA",family:"Nemotron",version:"340b",release_date:"2024-06-14",architecture:{type:"dense",total_params_b:340,active_params_b:340,num_layers:96,hidden_dim:18432,num_heads:96,num_kv_heads:8,head_dim:192,intermediate_dim:49152,vocab_size:256e3},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:680,weights_fp8_gb:340,weights_int4_gb:170,kv_per_token_bytes:2359296,activation_estimate_gb:8},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["tensorrt-llm","vllm","sglang"],api_pricing:{nvidia:{input_per_m:4.2,output_per_m:4.2}},tags:["nvidia","open-source","chat","code","open-weight","nvidia-nim-optimized"],huggingface_id:"nvidia/Nemotron-4-340B-Instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/nvidia/Nemotron-4-340B-Instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.185Z",license:"other"},{id:"nvidia/nemotron-70b",name:"Nemotron 70B",provider:"NVIDIA",family:"Nemotron",version:"1",release_date:"2024-10-14",architecture:{type:"dense",total_params_b:70.6,active_params_b:70.6,num_layers:80,hidden_dim:8192,num_heads:64,num_kv_heads:8,head_dim:128,intermediate_dim:28672,vocab_size:128256},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:14

11.2,weights_fp8_gb:70.6,weights_int4_gb:35.3,kv_per_token_bytes:327680,activation_estimate_gb:2.5},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm","sglang","tensorrt-llm"],api_pricing:{nvidia:{input_per_m:.88,output_per_m:.88}},tags:["chat","open-weight","reward-model-tuned","instruction-following"],huggingface_id:"nvidia/Llama-3.1-Nemotron-70B-Instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/nvidia/Llama-3.1-Nemotron-70B-Instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.186Z",license:"llama3.1"},{id:"nvidia/nemotron-mini-4b",name:"Nemotron Mini 4B",provider:"NVIDIA",family:"Nemotron",version:"1",release_date:"2024-08-01",architecture:{type:"dense",total_params_b:4,active_params_b:4,num_layers:32,hidden_dim:3072,num_heads:24,num_kv_heads:8,head_dim:128,intermediate_dim:9216,vocab_size:256e3},context_window:8192,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:8,weights_fp8_gb:4,weights_int4_gb:2,kv_per_token_bytes:49152,activation_estimate_gb:.5},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["tensorrt-llm","vllm","sglang"],api_pricing:{"nvidia-nim":{input_per_m:.06,output_per_m:.06}},tags:["nvidia","open-source","edge","optimized","nvidia-nim-optimized"],huggingface_id:"nvidia/Nemotron-Mini-4B-Instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/nvidia/Nemotron-Mini-4B-Instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.186Z",license:"other"},{id:"nvidia/nemotron-super-120b",name:"Nemotron-3 Super 120B",provider:"NVIDIA",family:"Nemotron",version:"3",release_date:"2025-03-18",architecture:{type:"dense",total_params_b:120,active_params_b:120,num_layers:80,hidden_dim:8192,num_heads:64,num_kv_heads:8,head_dim:128,intermediate_dim:28672,vocab_size:128256},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:240,weights_fp8_gb:120,weights_int4_gb:60,kv_per_token_bytes:327680,activation_estimate_gb:3.5},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm","sglang","tensorrt-llm"],api_pricing:{nvidia:{input_per_m:.8,output_per_m:2.4},openrouter:{input_per_m:.09,output_per_m:.45}},tags:["chat","code","reasoning","open-weight"],huggingface_id:"nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-NVFP4",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-NVFP4",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.186Z",license:"other"},{id:"nvidia/nemotron-ultra-253b",name:"Nemotron Ultra 253B",provider:"NVIDIA",family:"Nemotron",version:"1",release_date:"2025-03-18",architecture:{type:"dense",total_params_b:253,active_params_b:253,num_layers:96,hidden_dim:12288,num_heads:96,num_kv_heads:8,head_dim:128,intermediate_dim:49152,vocab_size:128256},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:506,weights_fp8_gb:253,weights_int4_gb:126.5,kv_per_token_bytes:393216,activation_estimate_gb:5},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm","tensorrt-llm"],api_pricing:{nvidia:{input_per_m:2,output_per_m:6}},tags:["chat","code","reasoning","open-weight"],huggingface_id:"nvidia/Llama-3_1-Nemotron-Ultra-253B-v1",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/nvidia/Llama-3_1-Nemotron-Ultra-253B-v1",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.186Z",license:"other"},{id:"nvidia/nv-clip",name:"NV-CLIP",provider:"NVIDIA",family:"NV-CLIP",version:"1.0",modality:"vision-embedding",release_date:"2024-03-18",release_date_source:"https://catalog.ngc.nvidia.com/orgs/nvidia/teams/nim/containers/nvclip",release_date_verified_at:"2026-06-25T16:00:00Z",last_verified_at:"2026-06-25T16:00:00Z",license:"nvidia-open-model-license",specialized_architecture:{type:"encoder",total_params_b:.428,num_layers:24,hidden_dim:1024,notes:"Dual-encoder CLIP variant — separate ViT-L/14 image encoder and 12-layer text transformer trained contrastively against ~5B image-text pairs (LAION-derived plus NVIDIA-curated). Produces 1024-dim aligned image and text embeddings suitable for cross-modal retrieval, zero-shot classification, and as a frozen feature extractor for downstream multimodal models. Image encoder accepts 224×224 RGB; text encoder accepts up to 77 BPE tokens."},specialized_memory:{weights_bf16_gb:.856,weights_fp8_gb:.428,activation_estimate_gb:.3},specialized_pricing:{kind:"per_embedding",input_per_m_chars:.04},description:"NV-CLIP is NVIDIA's tuned variant of OpenAI CLIP, packaged as an NVIDIA NIM container for production embedding workloads. Optimized on H100/L40S with TensorRT-LLM and INT4/FP8 quantization recipes.",intended_use:"Cross-modal retrieval, zero-shot image classification, multimodal RAG (image→text similarity), as a vision tower in downstream multimodal LLMs. Served via NIM.",tags:["vision","embeddings","clip","multimodal","retrieval","zero-shot","nim"]},{id:"nvidia/nv-dinov2",name:"DINOv2 ViT-g/14 (NVIDIA-optimized)",provider:"NVIDIA",family:"DINOv2",version:"2.0",modality:"vision-embedding",release_date:"2023-04-14",release_date_source:"https://huggingface.co/facebook/dinov2-giant",release_date_verified_at:"2026-06-25T16:00:00Z",last_verified_at:"2026-06-25T16:00:00Z",huggingface_id:"facebook/dinov2-giant",huggingface_id_status:"manual",license:"apache-2.0",specialized_architecture:{type:"encoder",total_params_b:1.1,num_layers:40,hidden_dim:1536,notes:"Self-supervised vision transformer (ViT-g/14). Trained without labels via DINO (self-distillation with no labels) v2 on the LVD-142M curated image corpus. 14×14 patch size, 40 transformer blocks. Produces per-patch + CLS token embeddings that transfer to classification, segmentation, depth, and retrieval without fine-tuning. Originally published by Meta FAIR;

1 NVIDIA serves an inference-optimized variant via NIM with FP8 quantization on H100."},specialized_memory:{weights_bf16_gb:2.2,weights_fp8_gb:1.1,activation_estimate_gb:.6},specialized_pricing:{kind:"free_open_weights"},description:"DINOv2 is a self-supervised vision foundation model — produces strong frozen image features that transfer to many downstream tasks without task-specific fine-tuning. The ViT-g/14 variant has 1.1B parameters and is currently the strongest open self-supervised vision encoder by published linear-probe benchmarks.",intended_use:"Frozen feature extraction for downstream vision tasks (classification, segmentation, depth estimation, instance retrieval), as a vision tower for vision-language models, or as a backbone for fine-tuning on small labeled datasets.",tags:["vision","self-supervised","embeddings","open-weight","vit","feature-extraction"]},{id:"nvidia/nv-embed-v2",name:"NV Embed v2",provider:"NVIDIA",family:"NV Embed",version:"2",release_date:"2024-09-18",architecture:{type:"dense",total_params_b:7.85,active_params_b:7.85,num_layers:32,hidden_dim:4096,num_heads:32,num_kv_heads:8,head_dim:128,intermediate_dim:14336,vocab_size:32e3},context_window:32768,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:15.7,weights_fp8_gb:7.85,weights_int4_gb:3.925,kv_per_token_bytes:131072,activation_estimate_gb:.8},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm"],api_pricing:{nvidia:{input_per_m:.012,output_per_m:.012}},tags:["embedding","open-weight"],huggingface_id:"nvidia/NV-Embed-v2",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/nvidia/NV-Embed-v2",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.186Z",license:"cc-by-nc-4.0"},{id:"nvidia/nv-embedqa-e5-v5",name:"NV EmbedQA E5 v5",provider:"NVIDIA",family:"NV EmbedQA",version:"5",release_date:"2024-06-01",architecture:{type:"dense",total_params_b:.33,active_params_b:.33,num_layers:24,hidden_dim:1024,num_heads:16,num_kv_heads:16,head_dim:64,intermediate_dim:4096,vocab_size:30522},context_window:512,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:.66,weights_fp8_gb:.33,weights_int4_gb:.17,kv_per_token_bytes:12288,activation_estimate_gb:.1},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["tensorrt-llm","vllm"],api_pricing:{"nvidia-nim":{input_per_m:.006,output_per_m:.006}},tags:["nvidia","open-source","embedding","qa","e5"],huggingface_id:"nvidia/NV-EmbedQA-E5-v5",huggingface_id_status:"manual",release_date_source:"https://huggingface.co/nvidia/NV-EmbedQA-E5-v5",last_verified_at:"2026-06-25T16:19:03.737Z"},{id:"nvidia/nv-embedqa-mistral-7b",name:"NV EmbedQA Mistral 7B",provider:"NVIDIA",family:"NV EmbedQA",version:"mistral-7b",release_date:"2024-06-01",architecture:{type:"dense",total_params_b:7.24,active_params_b:7.24,num_layers:32,hidden_dim:4096,num_heads:32,num_kv_heads:8,head_dim:128,intermediate_dim:14336,vocab_size:32e3},context_window:32768,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:14.48,weights_fp8_gb:7.24,weights_int4_gb:3.62,kv_per_token_bytes:65536,activation_estimate_gb:.7},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["tensorrt-llm","vllm","sglang"],api_pricing:{"nvidia-nim":{input_per_m:.012,output_per_m:.012}},tags:["nvidia","open-source","embedding","qa","mistral-based"],huggingface_id:"nvidia/NV-EmbedQA-Mistral7B-v2",huggingface_id_status:"manual",release_date_source:"https://huggingface.co/nvidia/NV-EmbedQA-Mistral7B-v2",last_verified_at:"2026-06-25T16:19:03.737Z"},{id:"nvidia/nv-retriever-v1",name:"NV Retriever v1",provider:"NVIDIA",family:"NV Retriever",version:"1",release_date:"2024-06-01",architecture:{type:"dense",total_params_b:.33,active_params_b:.33,num_layers:24,hidden_dim:1024,num_heads:16,num_kv_heads:16,head_dim:64,intermediate_dim:4096,vocab_size:30522},context_window:512,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:.66,weights_fp8_gb:.33,weights_int4_gb:.17,kv_per_token_bytes:12288,activation_estimate_gb:.1},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["tensorrt-llm","vllm"],api_pricing:{"nvidia-nim":{input_per_m:.006,output_per_m:.006}},tags:["nvidia","open-source","retrieval","embedding"],huggingface_id:"nvidia/NV-Retriever-v1",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/nvidia/NV-Retriever-v1",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.186Z",license:"other"},{id:"nvidia/nvlm-d-72b",name:"NVLM-D 72B",provider:"NVIDIA",family:"NVLM",version:"1",release_date:"2024-09-17",architecture:{type:"dense",total_params_b:72,active_params_b:72,num_layers:80,hidden_dim:8192,num_heads:64,num_kv_heads:8,head_dim:128,intermediate_dim:29568,vocab_size:152064},context_window:32768,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:144,weights_fp8_gb:72,weights_int4_gb:36,kv_per_token_bytes:327680,activation_estimate_gb:2.5},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm","tensorrt-llm"],api_pricing:{},tags:["vision","multimodal","open-weight"],huggingface_id:"nvidia/NVLM-D-72B",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/nvidia/NVLM-D-72B",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.186Z",license:"cc-by-nc-4.0"},{id:"nvidia/parakeet-ctc-0.6b",name:"Parakeet CTC 0.6B",provider:"NVIDIA",family:"Parakeet",version:"ctc",release_date:"2024-04-10",architecture:{type:"dense",total_params_b:.6,active_params_b:.6,num_layers:18,hidden_dim:768,num_heads:8,num_kv_heads:8,head_dim:96,intermediate_dim:3072,vocab_size:128e3},context_window:4096,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:1.2,weights_fp8_gb:.6,weights_int4_gb:.3,kv_per_token_bytes:6912,activation_estimate_gb:.1},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["tensorrt-llm","vllm"],api_pricing:{"nvidia-nim":{input_per_m:.03,output_per_m:.03}},tags:["nvidia","open-source","asr","speech-recognition","english"],huggingface_id:"nvidia/parakeet-ctc-0.6b",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/nvidia/parakeet-ctc-0.6b",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.186Z",license:"cc-by-4.0"},{id:"nvidia/parakeet-tdt-1.1b",name:"Parakeet TDT 1.1B",provider:"NVIDIA",family:"Parakeet",version:"tdt",release_date:"2024-04-10",architecture:{type:"dense",total_params_b:1.1,active_params_b:1.1,num_layers:24,hidden_dim:1024,num_heads:8,num_kv_heads:8,head_dim:128,intermediate_dim:4096,vocab_size:128e3},context_window:4096,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:2.2,weights_fp8_gb:1.1,weights_int4_gb:.55,kv_per_token_bytes:12288,activation_estimate_gb:.2},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["tensorrt-llm","vllm"],api_pricing:{"nvidia-nim":{input_per_m:.04,output_per_m:.04}},tags:["nvidia","open-source","asr","speech-recognition","english"],huggingface_id:"nvidia/parakeet-tdt-1.1b",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/nvidia/parakeet-tdt-1.1b",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.186Z",license:"cc-by-4.0"},{id:"nvidia/riva-fastpitch-en",name:"Riva FastPitch (en-US)",provider:"NVIDIA",family:"Riva TTS",version:"1.18",modality:"speech-tts",release_date:"2024-04-01",release_date_source:"https://catalog.ngc.nvidia.com/orgs/nvidia/teams/nemo/models/tts_en_fastpitch_ipa",release_date_verified_at:"2026-06-25T16:00:00Z",last_verified_at:"2026-06-25T16:00:00Z",huggingface_id:"nvidia/tts_en_fastpitch_ipa",huggingface_id_status:"manual",license:"cc-by-4.0",specialized_architecture:{type:"encoder-decoder",total_params_b:.045,num_layers:6,hidden_dim:384,notes:"Mel-spectrogram acoustic model. Transformer text encoder + duration/pitch predictors + transformer decoder predicting 80-band mel-spectrogram frames. Designed to pair with a separate vocoder (HiFi-GAN) that converts mel-spectrograms to waveform. Non-autoregressive — predicts all frames in parallel for sub-real-time inference."},specialized_memory:{weights_bf16_gb:.09,activation_estimate_gb:.05},specialized_pricing:{kind:"free_open_weights"},description:"FastPitch is a parallel transformer-based mel-spectrogram generator that explicitly controls pitch and duration of speech. Packaged 

1in NVIDIA NeMo / Riva as the front-end of the en-US neural TTS pipeline; the back-end vocoder is HiFi-GAN. Trained on the LJSpeech corpus.",intended_use:"Real-time text-to-speech for conversational AI, accessibility, voice interfaces. Pair with HiFi-GAN for full waveform synthesis.",tags:["tts","speech-synthesis","mel-spectrogram","open-weight","english","real-time"]},{id:"nvidia/riva-hifigan-en",name:"Riva HiFi-GAN (en-US)",provider:"NVIDIA",family:"Riva TTS",version:"1.18",modality:"speech-tts",release_date:"2024-04-01",release_date_source:"https://catalog.ngc.nvidia.com/orgs/nvidia/teams/nemo/models/tts_en_hifigan",release_date_verified_at:"2026-06-25T16:00:00Z",last_verified_at:"2026-06-25T16:00:00Z",huggingface_id:"nvidia/tts_hifigan",huggingface_id_status:"manual",license:"cc-by-4.0",specialized_architecture:{type:"gan",total_params_b:.014,notes:"GAN-based neural vocoder. Generator is a fully-convolutional decoder that upsamples 80-band mel-spectrograms to 22.05 kHz waveform via transposed convolutions + multi-receptive-field fusion (MRF) modules. Trained adversarially against a multi-period + multi-scale discriminator. Pairs with FastPitch (mel acoustic model) to complete the en-US TTS pipeline."},specialized_memory:{weights_bf16_gb:.028,activation_estimate_gb:.04},specialized_pricing:{kind:"free_open_weights"},description:"HiFi-GAN is the vocoder half of NVIDIA's en-US Riva TTS stack — converts mel-spectrograms produced by FastPitch into 22.05 kHz waveform audio. Known for high-fidelity sub-real-time synthesis on consumer GPUs.",intended_use:"Vocoder paired with a mel-spectrogram acoustic model (FastPitch, Tacotron 2) to produce waveform speech. Suitable for streaming TTS at well under 1× real-time.",tags:["tts","vocoder","gan","waveform-synthesis","open-weight","english","real-time"]}

1,{id:"nvidia/stylegan3",name:"StyleGAN3",provider:"NVIDIA",family:"StyleGAN",version:"3.0",modality:"image-gen",release_date:"2021-10-12",release_date_source:"https://nvlabs.github.io/stylegan3/",release_date_verified_at:"2026-06-25T16:00:00Z",last_verified_at:"2026-06-25T16:00:00Z",huggingface_id:"stabilityai/stylegan3",huggingface_id_status:"not_found",license:"nvidia-source-code-license-nc",specialized_architecture:{type:"gan",total_params_b:.025,notes:"Unconditional generative adversarial network with alias-free transformation layers. Generator is a fully-equivariant CNN that produces 1024×1024 RGB images from a 512-dim Gaussian latent code; discriminator is a CNN with progressive growing. Solved the texture-sticking and aliasing artifacts that plagued StyleGAN2 — outputs are translation- and rotation-equivariant under animation. Trained on FFHQ-U, AFHQv2, MetFaces-U corpora."},specialized_memory:{weights_bf16_gb:.05,activation_estimate_gb:.4},specialized_pricing:{kind:"free_open_weights"},description:"StyleGAN3 is NVIDIA Labs' third-generation generative adversarial network for photorealistic image synthesis. Introduced alias-free generator design that eliminated the rotation/translation artifacts present in prior StyleGAN versions. Research-grade — non-commercial license.",intended_use:"Research on generative image models, style transfer, latent-space manipulation. Pre-trained checkpoints for FFHQ-U (human faces), AFHQv2 (animal faces), MetFaces-U (art portraits). Inference on a single consumer GPU.",tags:["image-gen","gan","research","open-weight","non-commercial","alias-free"]},{id:"nvidia/vila-1.5-13b",name:"VILA 1.5 13B",provider:"NVIDIA",family:"VILA",version:"1.5",release_date:"2024-05-09",architecture:{type:"dense",total_params_b:13,active_params_b:13,num_layers:40,hidden_dim:5120,num_heads:40,num_kv_heads:8,head_dim:128,intermediate_dim:13824,vocab_size:32e3},context_window:4096,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:26,weights_fp8_gb:13,weights_int4_gb:6.5,kv_per_token_bytes:102400,activation_estimate_gb:1.2},capabilities:{tools:!1,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["tensorrt-llm","vllm","sglang"],api_pricing:{"nvidia-nim":{input_per_m:.3,output_per_m:.3}},tags:["nvidia","open-source","vision-language","multimodal"],huggingface_id:"Efficient-Large-Model/VILA1.5-13b",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/Efficient-Large-Model/VILA1.5-13b",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.186Z",license:"cc-by-nc-4.0"},{id:"nvidia/vila-1.5-3b",name:"VILA 1.5 3B",provider:"NVIDIA",family:"VILA",version:"1.5",release_date:"2024-05-09",architecture:{type:"dense",total_params_b:3,active_params_b:3,num_layers:26,hidden_dim:3200,num_heads:32,num_kv_heads:32,head_dim:100,intermediate_dim:8640,vocab_size:32e3},context_window:4096,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:6,weights_fp8_gb:3,weights_int4_gb:1.5,kv_per_token_bytes:41600,activation_estimate_gb:.4},capabilities:{tools:!1,vision:!0,structured_output:!1,code:!1,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["tensorrt-llm","vllm","sglang"],api_pricing:{"nvidia-nim":{input_per_m:.08,output_per_m:.08}},tags:["nvidia","open-source","vision-language","multimodal","edge"],huggingface_id:"Efficient-Large-Model/VILA1.5-3b",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/Efficient-Large-Model/VILA1.5-3b",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.186Z",license:"cc-by-nc-4.0"},{id:"nvidia/vila-1.5-40b",name:"VILA 1.5 40B",provider:"NVIDIA",family:"VILA",version:"1.5",release_date:"2024-05-09",architecture:{type:"dense",total_params_b:40,active_params_b:40,num_layers:48,hidden_dim:8192,num_heads:64,num_kv_heads:8,head_dim:128,intermediate_dim:28672,vocab_size:32e3},context_window:8192,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:80,weights_fp8_gb:40,weights_int4_gb:20,kv_per_token_bytes:196608,activation_estimate_gb:3},capabilities:{tools:!1,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["tensorrt-llm","vllm","sglang"],api_pricing:{"nvidia-nim":{input_per_m:1,output_per_m:1}},tags:["nvidia","open-source","vision-language","multimodal"],huggingface_id:"Efficient-Large-Model/VILA1.5-40b",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/Efficient-Large-Model/VILA1.5-40b",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.186Z",license:"cc-by-nc-4.0"},{id:"openai/dall-e-3",name:"DALL-E 3",provider:"OpenAI",family:"DALL-E",version:"3",release_date:"2023-10-03",architecture:{type:"dense",total_params_b:3.5,active_params_b:3.5,num_layers:32,hidden_dim:2560,num_heads:32,num_kv_heads:32,head_dim:80,intermediate_dim:10240,vocab_size:1e5},context_window:4096,supported_precisions:["bf16"],default_precision:"bf16",memory:{weights_bf16_gb:7,weights_fp8_gb:3.5,weights_int4_gb:1.75,kv_per_token_bytes:163840,activation_estimate_gb:1},capabilities:{tools:!1,vision:!0,structured_output:!1,code:!1,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:[],api_pricing:{openai:{input_per_m:40,output_per_m:40}},tags:["image-generation","proprietary","api-only","estimated-architecture"],huggingface_id_status:"n_a",release_date_source:"https://platform.openai.com/docs/models/dall-e-3",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-16T09:29:53.000Z"},{id:"openai/gpt-5.4-mini",name:"GPT-5.4 Mini",family:"GPT",provider:"OpenAI",is_partial:!0,release_date:"2026-03-17",release_date_source:"https://openrouter.ai/models/openai/gpt-5.4-mini",release_date_verified_at:"2026-06-20T16:00:00Z",last_verified_at:"2026-06-20T16:00:00Z",context_window:4e5,deployment:"api",api_pricing:{openai:{input_per_m:.75,output_per_m:4.5}},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},description:"GPT-5.4 Mini is OpenAI's cost-efficient mid-tier model — multimodal, tool-using and reasoning-capable with a 400K context window, at a fraction of flagship pricing."}

1,{id:"openai/gpt-5.4-nano",name:"GPT-5.4 Nano",family:"GPT",provider:"OpenAI",is_partial:!0,release_date:"2026-03-17",release_date_source:"https://openrouter.ai/models/openai/gpt-5.4-nano",release_date_verified_at:"2026-06-20T16:00:00Z",last_verified_at:"2026-06-20T16:00:00Z",context_window:4e5,deployment:"api",api_pricing:{openai:{input_per_m:.2,output_per_m:1.25}},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},description:"GPT-5.4 Nano is OpenAI's smallest, cheapest GPT-5.4 tier for high-volume, latency-sensitive workloads. Multimodal and tool-capable with a 400K context window."},{id:"openai/gpt-5-5-instant",name:"Gpt 5 5 Instant",family:"GPT",provider:"openai",is_partial:!0,release_date:"2026-05-05",release_date_source:"https://openai.com/index/gpt-5-5-instant",release_date_verified_at:"2026-05-15T16:22:38.040Z",last_verified_at:"2026-05-15T16:22:38.040Z"},{id:"openai/gpt-5.5-pro",name:"GPT-5.5 Pro",family:"GPT",provider:"OpenAI",is_partial:!0,release_date:"2026-04-24",release_date_source:"https://openrouter.ai/models/openai/gpt-5.5-pro",release_date_verified_at:"2026-06-20T16:00:00Z",last_verified_at:"2026-06-20T16:00:00Z",context_window:105e4,deployment:"api",api_pricing:{openai:{input_per_m:30,output_per_m:180}},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},description:"GPT-5.5 Pro is OpenAI's highest-capability tier of GPT-5.5, tuned for the hardest reasoning and agentic tasks. Multimodal, tool-using, ~1M-token context. Priced at a premium over standard GPT-5.5."},{id:"openai/gpt-5-5-with-trusted-access-for-cyber",name:"Gpt 5 5 With Trusted Access For Cyber",family:"GPT",provider:"openai",is_partial:!0,release_date:"2026-05-07",release_date_source:"https://openai.com/index/gpt-5-5-with-trusted-access-for-cyber",release_date_verified_at:"2026-05-15T16:22:13.547Z",last_verified_at:"2026-05-15T16:22:13.547Z"},{id:"openai/whisper-base",name:"Whisper Base",provider:"OpenAI",family:"Whisper",version:"1",release_date:"2022-09-21",architecture:{type:"dense",total_params_b:.074,active_params_b:.074,num_layers:6,hidden_dim:512,num_heads:8,num_kv_heads:8,head_dim:64,intermediate_dim:2048,vocab_size:51866},context_window:448,supported_precisions:["bf16","fp8"],default_precision:"bf16",memory:{weights_bf16_gb:.148,weights_fp8_gb:.074,weights_int4_gb:.037,kv_per_token_bytes:12288,activation_estimate_gb:.05},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","tensorrt-llm"],api_pricing:{},tags:["audio","transcription","open-weight","lightweight"],huggingface_id_status:"n_a",release_date_source:"https://platform.openai.com/docs/models/whisper-base",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-16T09:29:53.000Z"},{id:"openai/whisper-large-v3",name:"Whisper Large V3",provider:"OpenAI",family:"Whisper",version:"3",release_date:"2023-11-06",architecture:{type:"dense",total_params_b:1.55,active_params_b:1.55,num_layers:32,hidden_dim:1280,num_heads:20,num_kv_heads:20,head_dim:64,intermediate_dim:5120,vocab_size:51866},context_window:448,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:3.1,weights_fp8_gb:1.55,weights_int4_gb:.775,kv_per_token_bytes:163840,activation_estimate_gb:.3},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","tensorrt-llm"],api_pricing:{openai:{input_per_m:.006,output_per_m:.006}},tags:["audio","transcription","open-weight"],huggingface_id_status:"n_a",release_date_source:"https://platform.openai.com/docs/models/whisper-large-v3",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-16T09:29:53.000Z"},{id:"openai/whisper-medium",name:"Whisper Medium",provider:"OpenAI",family:"Whisper",version:"1",release_date:"2022-09-21",architecture:{type:"dense",total_params_b:.769,active_params_b:.769,num_layers:24,hidden_dim:1024,num_heads:16,num_kv_heads:16,head_dim:64,intermediate_dim:4096,vocab_size:51866},context_window:448,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:1.538,weights_fp8_gb:.769,weights_int4_gb:.385,kv_per_token_bytes:98304,activation_estimate_gb:.2},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","tensorrt-llm"],api_pricing:{},tags:["audio","transcription","open-weight"],huggingface_id_status:"n_a",release_date_source:"https://platform.openai.com/docs/models/whisper-medium",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-16T09:29:53.000Z"},{id:"openai/whisper-small",name:"Whisper Small",provider:"OpenAI",family:"Whisper",version:"1",release_date:"2022-09-21",architecture:{type:"dense",total_params_b:.244,active_params_b:.244,num_layers:12,hidden_dim:768,num_heads:12,num_kv_heads:12,head_dim:64,intermediate_dim:3072,vocab_size:51866},context_window:448,supported_precisions:["bf16","fp8"],default_precision:"bf16",memory:{weights_bf16_gb:.488,weights_fp8_gb:.244,weights_int4_gb:.122,kv_per_token_bytes:36864,activation_estimate_gb:.1},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","tensorrt-llm"],api_pricing:{},tags:["audio","transcription","open-weight","lightweight"],huggingface_id_status:"n_a",release_date_source:"https://platform.openai.com/docs/models/whisper-small",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-16T09:29:53.000Z"},{id:"proprietary/command-a",name:"Command A",provider:"Cohere",family:"Command",version:"a",release_date:"2025-03-13",architecture:{type:"dense",total_params_b:111,active_params_b:111,num_layers:72,hidden_dim:10240,num_heads:80,num_kv_heads:10,head_dim:128,intermediate_dim:36864,vocab_size:256e3},context_window:256e3,supported_precisions:["bf16"],default_precision:"bf16",memory:{weights_bf16_gb:222,weights_fp8_gb:111,weights_int4_gb:55.5,kv_per_token_bytes:184320,activation_estimate_gb:3.5},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:[],api_pricing:{cohere:{input_per_m:2.5,output_per_m:10},openrouter:{input_per_m:2.5,output_per_m:10}},tags:["proprietary","api-only","rag","enterprise"],huggingface_id:"CohereLabs/command-a-reasoning-08-2025",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/CohereLabs/command-a-reasoning-08-2025",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.186Z",license:"cc-by-nc-4.0"},{id:"proprietary/gemini-1.5-flash",name:"Gemini 1.5 Flash",provider:"Google",family:"Gemini",version:"1.5-flash",release_date:"2024-05-14",architecture:{type:"moe",total_params_b:50,active_params_b:12,num_layers:40,hidden_dim:5120,num_heads:40,num_kv_heads:8,head_dim:128,intermediate_dim:20480,vocab_size:256e3,num_experts:8,experts_per_token:2},context_window:1048576,supported_precisions:["bf16"],default_precision:"bf16",memory:{weights_bf16_gb:100,weights_fp8_gb:50,weights_int4_gb:25,kv_per_token_bytes:81920,activation_estimate_gb:2},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:[],api_pricing:{google:{input_per_m:.075,output_per_m:.3},openrouter:{input_per_m:.075,output_per_m:.3}},tags:["proprietary","api-only","estimated-architecture","fast","budget"],huggingface_id:"Safwanahmad619/gemini-1.5-flash-8b",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/Safwanahmad619/gemini-1.5-flash-8b",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-18T13:16:54.693Z"},{id:"proprietary/gemini-1.5-pro",name:"Gemini 1.5 Pro",provider:"Google",family:"Gemini",version:"1.5-pro",release_date:"2024-02-15",architecture:{type:"moe",total_params_b:175,active_params_b:40,num_layers:72,hidden_dim:8192,num_heads:64,num_kv_heads:8,head_dim:128,intermediate_dim:32768,vocab_size:256e3,num_experts:16,experts_per_token:2},context_window:2097152,supported_precisions:["bf16"],default_precision:"bf16",memory:{weights_bf16_gb:350,weights_fp8_gb:175,weights_int4_gb:87.5,kv_per_token_bytes:147456,activation_estimate_gb:4},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:[],api_pricing:{google:{input_per_m:1.25,output_per_m:5}},tags:["proprietary","api-only","estimated-architecture","long-context","multimodal"],release_date_source:"https://ai.google.dev/gemini-api/docs/models/gemini",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-16T09:29:53.000Z"},{id:"proprietary/gemini-2.0-flash",name:"Gemini 2.0 Flash",provider:"Google",family:"Gemini",version:"2.0-flash",release_date:"2025-02-05",architecture:{type:"moe",total_params_b:50,active_params_b:15,num_layers:48,hidden_dim:6144,num_heads:48,num_kv_heads:8,head_dim:128,intermediate_dim:24576,vocab_size:256e3,num_experts:8,experts_per_token:2},context_window:1048576,supported_precisions:["bf16"],default_precision:"bf16",memory:{weights_bf16_gb:100,weights_fp8_gb:50,weights_int4_gb:25,kv_per_token_bytes:98304,activation_estimate_gb:2},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:[],api_pricing:{google:{input_per_m:.1,output_per_m:.4},openrouter:{input_per_m:.1,output_per_m:.4}},tags:["proprietary","api-only","estimated-architecture","fast","multimodal"],release_date_source:"https://ai.google.dev/gemini-api/docs/models/gemini",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-16T09:29:53.000Z"},{id:"proprietary/gemini-2.0-pro",name:"Gemini 2.0 Pro",provider:"Google",family:"Gemini",version:"2.0-pro",release_date:"2025-02-05",architecture:{type:"moe",total_params_b:600,active_params_b:150,num_layers:96,hidden_dim:12288,num_heads:96,num_kv_heads:16,head_dim:128,intermediate_dim:49152,vocab_size:256e3,num_experts:16,experts_per_token:2},context_window:2e6,supported_precisions:["bf16"],default_precision:"bf16",memory:{weights_bf16_gb:1200,weights_fp8_gb:600,weights_int4_gb:300,kv_per_token_bytes:2359296,activation_estimate_gb:10},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:[],api_pricing:{google:{input_per_m:1,output_per_m:4}},tags:["proprietary","api-only","estimated-architecture"],release_date_source:"https://ai.google.dev/gemini-api/docs/models/gemini",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-16T09:29:53.000Z"},{id:"proprietary/gpt-3.5-turbo",name:"GPT-3.5 Turbo",provider:"OpenAI",family:"GPT-3.5",version:"3.5-turbo",release_date:"2023-11-06",architecture:{type:"dense",total_params_b:20,active_params_b:20,num_layers:40,hidden_dim:5120,num_heads:40,num_kv_heads:8,head_dim:128,intermediate_dim:13824,vocab_size:1e5},context_window:16384,supported_precisions:["bf16"],default_precision:"bf16",memory:{weights_bf16_gb:40,weights_fp8_gb:20,weights_int4_gb:10,kv_per_token_bytes:40960,activation_estimate_gb:1.5},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:[],api_pricing:{openai:{input_per_m:.5,output_per_m:1.5},openrouter:{input_per_m:.5,output_per_m:1.5}},tags:["proprietary","api-only","estimated-architecture","legacy","budget"],huggingface_id_status:"n_a",release_date_source:"https://platform.openai.com/docs/models/gpt-3.5-turbo",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-16T09:29:53.000Z"},{id:"proprietary/gpt-4-turbo",name:"GPT-4 Turbo",provider:"OpenAI",family:"GPT-4",version:"4-turbo",release_date:"2024-04-09",architecture:{type:"moe",total_params_b:200,active_params_b:50,num_layers:80,hidden_dim:10240,num_heads:80,num_kv_heads:10,head_dim:128,intermediate_dim:36864,vocab_size:1e5,num_experts:16,experts_per_token:2}

1,context_window:128e3,supported_precisions:["bf16"],default_precision:"bf16",memory:{weights_bf16_gb:400,weights_fp8_gb:200,weights_int4_gb:100,kv_per_token_bytes:204800,activation_estimate_gb:4},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:[],api_pricing:{openai:{input_per_m:10,output_per_m:30},openrouter:{input_per_m:30,output_per_m:60}},tags:["proprietary","api-only","estimated-architecture","legacy"],huggingface_id_status:"n_a",release_date_source:"https://platform.openai.com/docs/models/gpt-4-turbo",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-16T09:29:53.000Z"},{id:"proprietary/gpt-4.5-preview",name:"GPT-4.5 Preview",provider:"OpenAI",family:"GPT",version:"4.5-preview",release_date:"2025-02-27",architecture:{type:"moe",total_params_b:1500,active_params_b:300,num_layers:120,hidden_dim:16384,num_heads:128,num_kv_heads:16,head_dim:128,intermediate_dim:53248,vocab_size:2e5,num_experts:16,experts_per_token:2},context_window:128e3,supported_precisions:["bf16"],default_precision:"bf16",memory:{weights_bf16_gb:3e3,weights_fp8_gb:1500,weights_int4_gb:750,kv_per_token_bytes:3932160,activation_estimate_gb:20},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:[],api_pricing:{openai:{input_per_m:75,output_per_m:150}},tags:["proprietary","api-only","estimated-architecture"],huggingface_id_status:"n_a",release_date_source:"https://platform.openai.com/docs/models/gpt-4.5-preview",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-16T09:29:53.000Z"},{id:"proprietary/gpt-4o-mini",name:"GPT-4o Mini",provider:"OpenAI",family:"GPT-4",version:"4o-mini",release_date:"2024-07-18",architecture:{type:"dense",total_params_b:8,active_params_b:8,num_layers:32,hidden_dim:4096,num_heads:32,num_kv_heads:8,head_dim:128,intermediate_dim:14336,vocab_size:2e5},context_window:128e3,supported_precisions:["bf16"],default_precision:"bf16",memory:{weights_bf16_gb:16,weights_fp8_gb:8,weights_int4_gb:4,kv_per_token_bytes:16384,activation_estimate_gb:1},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:[],api_pricing:{openai:{input_per_m:.15,output_per_m:.6},openrouter:{input_per_m:.15,output_per_m:.6}},tags:["proprietary","api-only","estimated-architecture","fast","budget"],huggingface_id_status:"n_a",release_date_source:"https://platform.openai.com/docs/models/gpt-4o-mini",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-16T09:29:53.000Z"},{id:"proprietary/gpt-4o",name:"GPT-4o",provider:"OpenAI",family:"GPT-4",version:"4o",release_date:"2024-05-13",architecture:{type:"moe",total_params_b:200,active_params_b:50,num_layers:80,hidden_dim:10240,num_heads:80,num_kv_heads:10,head_dim:128,intermediate_dim:36864,vocab_size:2e5,num_experts:16,experts_per_token:2},context_window:128e3,supported_precisions:["bf16"],default_precision:"bf16",memory:{weights_bf16_gb:400,weights_fp8_gb:200,weights_int4_gb:100,kv_per_token_bytes:204800,activation_estimate_gb:4},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:[],api_pricing:{openai:{input_per_m:2.5,output_per_m:10},openrouter:{input_per_m:2.5,output_per_m:10},artificialanalysis:{input_per_m:2.5,output_per_m:10}},tags:["proprietary","api-only","estimated-architecture","flagship","multimodal"],huggingface_id_status:"n_a",release_date_source:"https://platform.openai.com/docs/models/gpt-4o",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-18T13:16:54.743Z"},{name:"GPT-5.5",provider:"OpenAI",family:"GPT",version:"5.5",release_date:"2026-04-24",release_date_source:"https://openrouter.ai/openai/gpt-5.5",release_date_verified_at:"2026-05-16T14:43:00Z",last_verified_at:"2026-05-16T14:43:00Z",context_window:1e6,supported_precisions:["bf16"],default_precision:"bf16",architecture:{type:"moe",total_params_b:700,active_params_b:110,num_layers:80,hidden_dim:10240,num_heads:80,num_kv_heads:10,head_dim:128,intermediate_dim:36864,vocab_size:2e5,num_experts:16,experts_per_token:2},memory:{weights_bf16_gb:1400,weights_fp8_gb:700,weights_int4_gb:350,kv_per_token_bytes:204800,activation_estimate_gb:4},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm"],api_pricing:{openai:{input_per_m:5,output_per_m:30,reasoning_per_m:30}},is_partial:!1,id:"proprietary/gpt-5-5"},{name:"GPT-5 Mini",provider:"OpenAI",family:"GPT",version:"5",release_date:"2025-08-0

17",release_date_source:"https://openrouter.ai/openai/gpt-5-mini",release_date_verified_at:"2026-05-16T14:43:00Z",last_verified_at:"2026-05-16T14:43:00Z",context_window:4e5,supported_precisions:["bf16"],default_precision:"bf16",architecture:{type:"moe",total_params_b:80,active_params_b:25,num_layers:80,hidden_dim:10240,num_heads:80,num_kv_heads:10,head_dim:128,intermediate_dim:36864,vocab_size:2e5,num_experts:16,experts_per_token:2},memory:{weights_bf16_gb:160,weights_fp8_gb:80,weights_int4_gb:40,kv_per_token_bytes:204800,activation_estimate_gb:4},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm"],api_pricing:{openai:{input_per_m:.25,output_per_m:2,reasoning_per_m:2}},is_partial:!1,id:"proprietary/gpt-5-mini"},{name:"GPT-5 Nano",provider:"OpenAI",family:"GPT",version:"5",release_date:"2025-08-07",release_date_source:"https://openrouter.ai/openai/gpt-5-nano",release_date_verified_at:"2026-05-16T14:43:00Z",last_verified_at:"2026-05-16T14:43:00Z",context_window:4e5,supported_precisions:["bf16"],default_precision:"bf16",architecture:{type:"moe",total_params_b:8,active_params_b:4,num_layers:80,hidden_dim:10240,num_heads:80,num_kv_heads:10,head_dim:128,intermediate_dim:36864,vocab_size:2e5,num_experts:16,experts_per_token:2},memory:{weights_bf16_gb:16,weights_fp8_gb:8,weights_int4_gb:4,kv_per_token_bytes:204800,activation_estimate_gb:4},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm"],api_pricing:{openai:{input_per_m:.05,output_per_m:.4,reasoning_per_m:.4}},is_partial:!1,id:"proprietary/gpt-5-nano"},{id:"proprietary/gpt-5",name:"GPT-5",provider:"OpenAI",family:"GPT",version:"5",release_date:"2025-08-07",release_date_source:"https://openrouter.ai/openai/gpt-5",release_date_verified_at:"2026-05-16T14:20:26Z",last_verified_at:"2026-05-16T14:20:26Z",is_partial:!1,tags:["frontier","reasoning"],context_window:4e5,supported_precisions:["bf16"],default_precision:"bf16",architecture:{type:"moe",total_params_b:500,active_params_b:90,num_layers:120,hidden_dim:12288,num_heads:96,num_kv_heads:10,head_dim:128,intermediate_dim:36864,vocab_size:2e5,num_experts:16,experts_per_token:2},memory:{weights_bf16_gb:1e3,weights_fp8_gb:500,weights_int4_gb:250,kv_per_token_bytes:204800,activation_estimate_gb:4},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm"],api_pricing:{openai:{input_per_m:1.25,output_per_m:10,reasoning_per_m:10},openrouter:{input_per_m:1.25,output_per_m:10}}},{id:"proprietary/grok-3",name:"Grok 3",provider:"xAI",family:"Grok",version:"3",release_date:"2025-02-18",architecture:{type:"moe",total_params_b:600,active_params_b:120,num_layers:96,hidden_dim:12288,num_heads:96,num_kv_heads:16,head_dim:128,intermediate_dim:49152,vocab_size:131072,num_experts:16,experts_per_token:2},context_window:131072,supported_precisions:["bf16"],default_precision:"bf16",memory:{weights_bf16_gb:1200,weights_fp8_gb:600,weights_int4_gb:300,kv_per_token_bytes:2359296,activation_estimate_gb:10},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:[],api_pricing:{xai:{input_per_m:3,output_per_m:15},openrouter:{input_per_m:3,output_per_m:15}},tags:["proprietary","api-only","estimated-architecture"],huggingface_id_status:"n_a",release_date_source:"https://x.ai/",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-16T09:29:53.000Z"},{id:"proprietary/inflection-3",name:"Inflection 3",provider:"Inflection AI",family:"Inflection",version:"3",release_date:"2024-03-07",architecture:{type:"dense",total_params_b:100,active_params_b:100,num_layers:72,hidden_dim:10240,num_heads:80,num_kv_heads:10,head_dim:128,intermediate_dim:36864,vocab_size:128e3},context_window:8192,supported_precisions:["bf16"],default_precision:"bf16",memory:{weights_bf16_gb:200,weights_fp8_gb:100,weights_int4_gb:50,kv_per_token_bytes:184320,activation_estimate_gb:3.5},capabilities:{tools:!1,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:[],api_pricing:{inflection:{input_per_m:5,output_per_m:15},openrouter:{input_per_m:2.5,output_per_m:10}},tags:["proprietary","api-only","estimated-architecture","conversational"],release_date_source:"https://inflection.ai/blog/inflection-3",license:"proprietary",release_date_verified_at:"2026-06-25T16:20:15.668Z",last_verified_at:"2026-06-25T16:20:15.668Z"}

1,{id:"proprietary/jamba-instruct",name:"Jamba Instruct",provider:"AI21",family:"Jamba",version:"instruct",release_date:"2024-03-28",architecture:{type:"moe",total_params_b:52,active_params_b:12,num_layers:32,hidden_dim:4096,num_heads:32,num_kv_heads:8,head_dim:128,intermediate_dim:14336,vocab_size:65536,num_experts:16,experts_per_token:2},context_window:256e3,supported_precisions:["bf16"],default_precision:"bf16",memory:{weights_bf16_gb:104,weights_fp8_gb:52,weights_int4_gb:26,kv_per_token_bytes:65536,activation_estimate_gb:2},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:[],api_pricing:{ai21:{input_per_m:.5,output_per_m:.7}},tags:["proprietary","api-only","mamba","hybrid","long-context"],huggingface_id:"FreedomIntelligence/Jamba-9B-Instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/FreedomIntelligence/Jamba-9B-Instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.186Z",license:"mit"},{id:"proprietary/llama-4-behemoth",name:"Llama 4 Behemoth",provider:"Meta",family:"Llama 4",version:"4-behemoth",release_date:"2025-04-05",release_date_source:"https://ai.meta.com/blog/llama-4-multimodal-intelligence/",release_date_verified_at:"2026-05-13T14:42:10Z",last_verified_at:"2026-05-13T14:42:10Z",architecture:{type:"moe",total_params_b:2e3,active_params_b:400,num_layers:128,hidden_dim:16384,num_heads:128,num_kv_heads:16,head_dim:128,intermediate_dim:53248,vocab_size:202400,num_experts:256,experts_per_token:16},context_window:1048576,supported_precisions:["bf16"],default_precision:"bf16",memory:{weights_bf16_gb:4e3,weights_fp8_gb:2e3,weights_int4_gb:1e3,kv_per_token_bytes:4194304,activation_estimate_gb:25},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:[],api_pricing:{together:{input_per_m:5,output_per_m:16}},tags:["proprietary","api-only","estimated-architecture","moe"]},{id:"proprietary/o1-mini",name:"o1-mini",provider:"OpenAI",family:"o1",version:"mini",release_date:"2024-09-12",architecture:{type:"dense",total_params_b:70,active_params_b:70,num_layers:64,hidden_dim:8192,num_heads:64,num_kv_heads:8,head_dim:128,intermediate_dim:28672,vocab_size:2e5},context_window:128e3,supported_precisions:["bf16"],default_precision:"bf16",memory:{weights_bf16_gb:140,weights_fp8_gb:70,weights_int4_gb:35,kv_per_token_bytes:131072,activation_estimate_gb:3},capabilities:{tools:!1,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:[],api_pricing:{openai:{input_per_m:3,output_per_m:12,reasoning_per_m:12}},tags:["proprietary","api-only","estimated-architecture","reasoning"],huggingface_id_status:"n_a",release_date_source:"https://platform.openai.com/docs/models/o1-mini",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-16T09:29:53.000Z"},{id:"proprietary/o1",name:"o1",provider:"OpenAI",family:"o1",version:"1",release_date:"2024-12-17",architecture:{type:"moe",total_params_b:200,active_params_b:50,num_layers:80,hidden_dim:10240,num_heads:80,num_kv_heads:10,head_dim:128,intermediate_dim:36864,vocab_size:2e5,num_experts:16,experts_per_token:2},context_window:2e5,supported_precisions:["bf16"],default_precision:"bf16",memory:{weights_bf16_gb:400,weights_fp8_gb:200,weights_int4_gb:100,kv_per_token_bytes:204800,activation_estimate_gb:4},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:[],api_pricing:{openai:{input_per_m:15,output_per_m:60,reasoning_per_m:60},openrouter:{input_per_m:15,output_per_m:60}},tags:["proprietary","api-only","estimated-architecture","reasoning","flagship"],huggingface_id_status:"n_a",release_date_source:"https://platform.openai.com/docs/models/o1",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-16T09:29:53.000Z"},{id:"proprietary/o3-mini",name:"o3-mini",provider:"OpenAI",family:"o3",version:"mini",release_date:"2025-01-31",architecture:{type:"dense",total_params_b:70,active_params_b:70,num_layers:64,hidden_dim:8192,num_heads:64,num_kv_heads:8,head_dim:128,intermediate_dim:28672,vocab_size:2e5}

1,context_window:2e5,supported_precisions:["bf16"],default_precision:"bf16",memory:{weights_bf16_gb:140,weights_fp8_gb:70,weights_int4_gb:35,kv_per_token_bytes:131072,activation_estimate_gb:3},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:[],api_pricing:{openai:{input_per_m:1.1,output_per_m:4.4,reasoning_per_m:4.4},openrouter:{input_per_m:1.1,output_per_m:4.4}},tags:["proprietary","api-only","estimated-architecture","reasoning"],huggingface_id_status:"n_a",release_date_source:"https://platform.openai.com/docs/models/o3-mini",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-16T09:29:53.000Z"},{id:"proprietary/reka-core",name:"Reka Core",provider:"Reka AI",family:"Reka",version:"core",release_date:"2024-04-15",architecture:{type:"dense",total_params_b:70,active_params_b:70,num_layers:64,hidden_dim:8192,num_heads:64,num_kv_heads:8,head_dim:128,intermediate_dim:28672,vocab_size:128e3},context_window:128e3,supported_precisions:["bf16"],default_precision:"bf16",memory:{weights_bf16_gb:140,weights_fp8_gb:70,weights_int4_gb:35,kv_per_token_bytes:131072,activation_estimate_gb:3},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:[],api_pricing:{reka:{input_per_m:3,output_per_m:15}},tags:["proprietary","api-only","estimated-architecture","multimodal"],release_date_source:"https://www.reka.ai/news/reka-core-our-frontier-class-multimodal-language-model",license:"proprietary",release_date_verified_at:"2026-06-25T16:20:15.669Z",last_verified_at:"2026-06-25T16:20:15.669Z"},{id:"qwen/qwen-2-vl-2b",name:"Qwen 2 VL 2B",provider:"Alibaba",family:"Qwen 2 VL",version:"2",release_date:"2024-08-29",architecture:{type:"dense",total_params_b:2.2,active_params_b:2.2,num_layers:28,hidden_dim:1536,num_heads:12,num_kv_heads:2,head_dim:128,intermediate_dim:8960,vocab_size:151936},context_window:32768,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:4.4,weights_fp8_gb:2.2,weights_int4_gb:1.1,kv_per_token_bytes:14336,activation_estimate_gb:.3},capabilities:{tools:!1,vision:!0,structured_output:!0,code:!1,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi"],api_pricing:{},tags:["vision","multimodal","open-weight","lightweight"],huggingface_id:"Qwen/Qwen3-VL-2B-Instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/Qwen/Qwen3-VL-2B-Instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.186Z",license:"apache-2.0"},{id:"qwen/qwen-2.5-0.5b",name:"Qwen 2.5 0.5B",provider:"Alibaba",family:"Qwen 2.5",version:"2.5",release_date:"2024-09-19",architecture:{type:"dense",total_params_b:.5,active_params_b:.5,num_layers:24,hidden_dim:896,num_heads:14,num_kv_heads:2,head_dim:64,intermediate_dim:4864,vocab_size:151936},context_window:32768,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:1,weights_fp8_gb:.5,weights_int4_gb:.25,kv_per_token_bytes:12288,activation_estimate_gb:.2},capabilities:{tools:!1,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","ollama"],api_pricing:{},tags:["chat","code","open-weight","lightweight"],huggingface_id:"Qwen/Qwen2.5-0.5B-Instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/Qwen/Qwen2.5-0.5B-Instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.186Z",license:"apache-2.0"},{id:"qwen/qwen-2.5-1.5b",name:"Qwen 2.5 1.5B",provider:"Alibaba",family:"Qwen 2.5",version:"2.5",release_date:"2024-09-19",architecture:{type:"dense",total_params_b:1.5,active_params_b:1.5,num_layers:28,hidden_dim:1536,num_heads:12,num_kv_heads:2,head_dim:128,intermediate_dim:8960,vocab_size:151936},context_window:32768,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:3,weights_fp8_gb:1.5,weights_int4_gb:.75,kv_per_token_bytes:28672,activation_estimate_gb:.3},capabilities:{tools:!1,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","ollama"],api_pricing:{},tags:["chat","code","open-weight","lightweight"],huggingface_id:"Qwen

1/Qwen2.5-1.5B-Instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/Qwen/Qwen2.5-1.5B-Instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.186Z",license:"apache-2.0"},{id:"qwen/qwen-2.5-72b",name:"Qwen 2.5 72B",provider:"Alibaba",family:"Qwen 2.5",version:"2.5",release_date:"2024-09-19",architecture:{type:"dense",total_params_b:72.7,active_params_b:72.7,num_layers:80,hidden_dim:8192,num_heads:64,num_kv_heads:8,head_dim:128,intermediate_dim:29568,vocab_size:152064},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:145.4,weights_fp8_gb:72.7,weights_int4_gb:36.35,kv_per_token_bytes:327680,activation_estimate_gb:2.5},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm"],api_pricing:{together:{input_per_m:1.2,output_per_m:1.2},fireworks:{input_per_m:.9,output_per_m:.9},novita:{input_per_m:.38,output_per_m:.4},featherless:{input_per_m:0,output_per_m:0},openrouter:{input_per_m:.36,output_per_m:.4},deepinfra:{input_per_m:.35,output_per_m:.4}},tags:["chat","code","instruction-following","open-weight"],huggingface_id:"Qwen/Qwen2.5-72B-Instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/Qwen/Qwen2.5-72B-Instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.186Z",license:"other"},{id:"qwen/qwen-2.5-7b",name:"Qwen 2.5 7B",provider:"Alibaba",family:"Qwen 2.5",version:"2.5",release_date:"2024-09-19",architecture:{type:"dense",total_params_b:7.6,active_params_b:7.6,num_layers:28,hidden_dim:3584,num_heads:28,num_kv_heads:4,head_dim:128,intermediate_dim:18944,vocab_size:152064},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:15.2,weights_fp8_gb:7.6,weights_int4_gb:3.8,kv_per_token_bytes:57344,activation_estimate_gb:1},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{together:{input_per_m:.2,output_per_m:.2},fireworks:{input_per_m:.2,output_per_m:.2},featherless:{input_per_m:0,output_per_m:0},openrouter:{input_per_m:.04,output_per_m:.1}},tags:["chat","code","instruction-following","open-weight"],huggingface_id:"Qwen/Qwen2.5-7B-Instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/Qwen/Qwen2.5-7B-Instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.186Z",license:"apache-2.0"},{id:"qwen/qwen-2.5-coder-1.5b",name:"Qwen 2.5 Coder 1.5B",provider:"Alibaba",family:"Qwen 2.5 Coder",version:"2.5",release_date:"2024-11-12",architecture:{type:"dense",total_params_b:1.5,active_params_b:1.5,num_layers:28,hidden_dim:1536,num_heads:12,num_kv_heads:2,head_dim:128,intermediate_dim:8960,vocab_size:152064},context_window:32768,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:3,weights_fp8_gb:1.5,weights_int4_gb:.75,kv_per_token_bytes:14336,activation_estimate_gb:.2},capabilities:{tools:!1,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","ollama"],api_pricing:{},tags:["code","code-specialized","open-weight","lightweight"],huggingface_id:"Qwen/Qwen2.5-Coder-1.5B",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/Qwen/Qwen2.5-Coder-1.5B",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.186Z",license:"apache-2.0"},{id:"qwen/qwen-2.5-coder-14b",name:"Qwen 2.5 Coder 14B",provider:"Alibaba",family:"Qwen 2.5 Coder",version:"2.5",release_date:"2024-11-12",architecture:{type:"dense",total_params_b:14.7,active_params_b:14.7,num_layers:48,hidden_dim:5120,num_heads:40,num_kv_heads:8,head_dim:128,intermediate_dim:13696,vocab_size:152064},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:29.4,weights_fp8_gb:14.7,weights_int4_gb:7.35,kv_per_token_bytes:196608,activation_estimate_gb:1.2},capabilities:{tools:!1,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{together:{input_per_m:.3,output_per_m:.3}},tags:["code","code-specialized","open-weight"],huggingface_id:"Qwen

1/Qwen2.5-Coder-14B-Instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/Qwen/Qwen2.5-Coder-14B-Instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.187Z",license:"apache-2.0"},{id:"qwen/qwen-2.5-coder-32b",name:"Qwen 2.5 Coder 32B",provider:"Alibaba",family:"Qwen 2.5 Coder",version:"2.5",release_date:"2024-11-12",architecture:{type:"dense",total_params_b:32.5,active_params_b:32.5,num_layers:64,hidden_dim:5120,num_heads:40,num_kv_heads:8,head_dim:128,intermediate_dim:27648,vocab_size:152064},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:65,weights_fp8_gb:32.5,weights_int4_gb:16.25,kv_per_token_bytes:262144,activation_estimate_gb:1.8},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{together:{input_per_m:.8,output_per_m:.8},featherless:{input_per_m:0,output_per_m:0},scaleway:{input_per_m:0,output_per_m:0},openrouter:{input_per_m:.66,output_per_m:1}},tags:["code","chat","open-weight"],huggingface_id:"Qwen/Qwen2.5-Coder-32B-Instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/Qwen/Qwen2.5-Coder-32B-Instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.187Z",license:"apache-2.0"},{id:"qwen/qwen-2.5-coder-3b",name:"Qwen 2.5 Coder 3B",provider:"Alibaba",family:"Qwen 2.5 Coder",version:"2.5",release_date:"2024-11-12",architecture:{type:"dense",total_params_b:3.1,active_params_b:3.1,num_layers:36,hidden_dim:2048,num_heads:16,num_kv_heads:2,head_dim:128,intermediate_dim:11008,vocab_size:152064},context_window:32768,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:6.2,weights_fp8_gb:3.1,weights_int4_gb:1.55,kv_per_token_bytes:18432,activation_estimate_gb:.4},capabilities:{tools:!1,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","ollama"],api_pricing:{},tags:["code","code-specialized","open-weight","lightweight"],huggingface_id:"Qwen/Qwen2.5-Coder-3B-Instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/Qwen/Qwen2.5-Coder-3B-Instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.187Z",license:"other"},{id:"qwen/qwen-2.5-coder-7b",name:"Qwen 2.5 Coder 7B",provider:"Alibaba",family:"Qwen 2.5 Coder",version:"2.5",release_date:"2024-11-12",architecture:{type:"dense",total_params_b:7.6,active_params_b:7.6,num_layers:28,hidden_dim:3584,num_heads:28,num_kv_heads:4,head_dim:128,intermediate_dim:18944,vocab_size:152064},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:15.2,weights_fp8_gb:7.6,weights_int4_gb:3.8,kv_per_token_bytes:57344,activation_estimate_gb:.8},capabilities:{tools:!1,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{together:{input_per_m:.2,output_per_m:.2}},tags:["code","code-specialized","open-weight"],huggingface_id:"Qwen/Qwen2.5-Coder-7B-Instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/Qwen/Qwen2.5-Coder-7B-Instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.187Z",license:"apache-2.0"},{id:"qwen/qwen-2.5-math-72b",name:"Qwen 2.5 Math 72B",provider:"Alibaba",family:"Qwen 2.5 Math",version:"2.5",release_date:"2024-09-19",architecture:{type:"dense",total_params_b:72.7,active_params_b:72.7,num_layers:80,hidden_dim:8192,num_heads:64,num_kv_heads:8,head_dim:128,intermediate_dim:29568,vocab_size:152064},context_window:4096,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:145.4,weights_fp8_gb:72.7,weights_int4_gb:36.35,kv_per_token_bytes:327680,activation_estimate_gb:2.5},capabilities:{tools:!1,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!1,reasoning:!0},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm"],api_pricing:{together:{input_per_m:.9,output_per_m:.9}},tags:["math","math-specialized","reasoning","open-weight"],huggingface_id:"Qwen

1/Qwen2.5-Math-72B-Instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/Qwen/Qwen2.5-Math-72B-Instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.187Z",license:"other"},{id:"qwen/qwen-2.5-math-7b",name:"Qwen 2.5 Math 7B",provider:"Alibaba",family:"Qwen 2.5 Math",version:"2.5",release_date:"2024-09-19",architecture:{type:"dense",total_params_b:7.6,active_params_b:7.6,num_layers:28,hidden_dim:3584,num_heads:28,num_kv_heads:4,head_dim:128,intermediate_dim:18944,vocab_size:152064},context_window:4096,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:15.2,weights_fp8_gb:7.6,weights_int4_gb:3.8,kv_per_token_bytes:57344,activation_estimate_gb:.8},capabilities:{tools:!1,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!1,reasoning:!0},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{together:{input_per_m:.2,output_per_m:.2}},tags:["math","math-specialized","reasoning","open-weight"],huggingface_id:"Qwen/Qwen2.5-Math-7B",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/Qwen/Qwen2.5-Math-7B",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.187Z",license:"apache-2.0"},{id:"qwen/qwen-2.5-vl-72b",name:"Qwen 2.5 VL 72B",provider:"Alibaba",family:"Qwen 2.5 VL",version:"2.5",release_date:"2025-01-27",architecture:{type:"dense",total_params_b:72.7,active_params_b:72.7,num_layers:80,hidden_dim:8192,num_heads:64,num_kv_heads:8,head_dim:128,intermediate_dim:29568,vocab_size:152064},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:145.4,weights_fp8_gb:72.7,weights_int4_gb:36.35,kv_per_token_bytes:327680,activation_estimate_gb:3},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm"],api_pricing:{together:{input_per_m:.9,output_per_m:.9},ovhcloud:{input_per_m:1.01,output_per_m:1.01},openrouter:{input_per_m:.25,output_per_m:.75}},tags:["chat","vision","vision-language","code","open-weight"],huggingface_id:"Qwen/Qwen2.5-VL-72B-Instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/Qwen/Qwen2.5-VL-72B-Instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.187Z",license:"other"},{id:"qwen/qwen-2.5-vl-7b",name:"Qwen 2.5 VL 7B",provider:"Alibaba",family:"Qwen 2.5 VL",version:"2.5",release_date:"2025-01-27",architecture:{type:"dense",total_params_b:7.6,active_params_b:7.6,num_layers:28,hidden_dim:3584,num_heads:28,num_kv_heads:4,head_dim:128,intermediate_dim:18944,vocab_size:152064},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:15.2,weights_fp8_gb:7.6,weights_int4_gb:3.8,kv_per_token_bytes:57344,activation_estimate_gb:.8},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{together:{input_per_m:.2,output_per_m:.2}},tags:["chat","vision","vision-language","code","open-weight"],huggingface_id:"Qwen/Qwen2.5-VL-7B-Instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/Qwen/Qwen2.5-VL-7B-Instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.187Z",license:"apache-2.0"},{id:"qwen/qwen-3-0.6b",name:"Qwen 3 0.6B",provider:"Alibaba",family:"Qwen 3",version:"3",release_date:"2025-04-28",architecture:{type:"dense",total_params_b:.6,active_params_b:.6,num_layers:28,hidden_dim:1024,num_heads:16,num_kv_heads:8,head_dim:128,intermediate_dim:3072,vocab_size:151936},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:1.2,weights_fp8_gb:.6,weights_int4_gb:.3,kv_per_token_bytes:57344,activation_estimate_gb:.2},capabilities:{tools:!1,vision:!1,structured_output:!0,code:!0,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{},tags:["chat","open-weight","lightweight","estimated-architecture"],huggingface_id:"Qwen

1/Qwen3-0.6B",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/Qwen/Qwen3-0.6B",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.187Z",license:"apache-2.0"},{id:"qwen/qwen-3-1.7b",name:"Qwen 3 1.7B",provider:"Alibaba",family:"Qwen 3",version:"3",release_date:"2025-04-28",architecture:{type:"dense",total_params_b:1.7,active_params_b:1.7,num_layers:28,hidden_dim:1536,num_heads:16,num_kv_heads:8,head_dim:128,intermediate_dim:8960,vocab_size:151936},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:3.4,weights_fp8_gb:1.7,weights_int4_gb:.85,kv_per_token_bytes:57344,activation_estimate_gb:.3},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{},tags:["chat","code","reasoning","open-weight","lightweight","estimated-architecture"],huggingface_id:"Qwen/Qwen3-1.7B",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/Qwen/Qwen3-1.7B",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.187Z",license:"apache-2.0"},{id:"qwen/qwen-3-235b",name:"Qwen 3 235B",provider:"Alibaba",family:"Qwen 3",version:"3",release_date:"2025-04-28",architecture:{type:"moe",total_params_b:235,active_params_b:22,num_layers:94,hidden_dim:5120,num_heads:64,num_kv_heads:4,head_dim:128,intermediate_dim:13824,vocab_size:151936,num_experts:128,experts_per_token:8},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:470,weights_fp8_gb:235,weights_int4_gb:117.5,kv_per_token_bytes:192512,activation_estimate_gb:3},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm","sglang","tensorrt-llm"],api_pricing:{together:{input_per_m:1.5,output_per_m:3},fireworks:{input_per_m:1.8,output_per_m:3.5},novita:{input_per_m:.2,output_per_m:.8},featherless:{input_per_m:0,output_per_m:0},openrouter:{input_per_m:.455,output_per_m:1.82}},tags:["chat","code","reasoning","moe","open-weight","flagship"],huggingface_id:"Qwen/Qwen3-235B-A22B",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/Qwen/Qwen3-235B-A22B",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.187Z",license:"apache-2.0"},{id:"qwen/qwen-3-30b-a3b",name:"Qwen 3 30B-A3B",provider:"Alibaba",family:"Qwen 3",version:"3",release_date:"2025-04-29",architecture:{type:"moe",total_params_b:30.5,active_params_b:3.3,num_layers:48,hidden_dim:2048,num_heads:32,num_kv_heads:4,head_dim:128,intermediate_dim:8192,vocab_size:151936,num_experts:128,experts_per_token:8},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:61,weights_fp8_gb:30.5,weights_int4_gb:15.25,kv_per_token_bytes:24576,activation_estimate_gb:.5},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm","sglang","tgi","ollama"],api_pricing:{novita:{input_per_m:.09,output_per_m:.45},openrouter:{input_per_m:.09,output_per_m:.45}},tags:["chat","code","moe","reasoning","open-weight"],huggingface_id:"Qwen/Qwen3-30B-A3B",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/Qwen/Qwen3-30B-A3B",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.187Z",license:"apache-2.0"},{id:"qwen/qwen-3-32b",name:"Qwen 3 32B",provider:"Alibaba",family:"Qwen 3",version:"3",release_date:"2025-04-28",architecture:{type:"dense",total_params_b:32.8,active_params_b:32.8,num_layers:64,hidden_dim:5120,num_heads:40,num_kv_heads:8,head_dim:128,intermediate_dim:25600,vocab_size:151936},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:65.6,weights_fp8_gb:32.8,weights_int4_gb:16.4,kv_per_token_bytes:262144,activation_estimate_gb:2},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{together:{input_per_m:.8,output_per_m:.8},fireworks:{input_per_m:.9,output_per_m:.9},novita:{input_per_m:.1,output_per_m:.45},featherless:{input_per_m:0,output_per_m:0},groq:{input_per_m:.29,output_per_m:.59},ovhcloud:{input_per_m:.09,output_per_m:.25},openrouter:{input_per_m:.08,output_per_m:.28}},tags:["chat","code","reasoning","open-weight"],huggingface_id:"Qwen

1/Qwen3-32B",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/Qwen/Qwen3-32B",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.187Z",license:"apache-2.0"},{id:"qwen/qwen-3-4b",name:"Qwen 3 4B",provider:"Alibaba",family:"Qwen 3",version:"3",release_date:"2025-04-28",architecture:{type:"dense",total_params_b:4,active_params_b:4,num_layers:36,hidden_dim:2560,num_heads:32,num_kv_heads:8,head_dim:128,intermediate_dim:9216,vocab_size:151936},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:8,weights_fp8_gb:4,weights_int4_gb:2,kv_per_token_bytes:147456,activation_estimate_gb:.5},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{together:{input_per_m:.1,output_per_m:.1}},tags:["chat","code","reasoning","open-weight","estimated-architecture"],huggingface_id:"Qwen/Qwen3-4B-Instruct-2507",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/Qwen/Qwen3-4B-Instruct-2507",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.187Z",license:"apache-2.0"},{id:"qwen/qwen-3-8b",name:"Qwen 3 8B",provider:"Alibaba",family:"Qwen 3",version:"3",release_date:"2025-04-28",architecture:{type:"dense",total_params_b:8.2,active_params_b:8.2,num_layers:36,hidden_dim:4096,num_heads:32,num_kv_heads:8,head_dim:128,intermediate_dim:14336,vocab_size:151936},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:16.4,weights_fp8_gb:8.2,weights_int4_gb:4.1,kv_per_token_bytes:147456,activation_estimate_gb:1},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{together:{input_per_m:.2,output_per_m:.2},fireworks:{input_per_m:.2,output_per_m:.2},featherless:{input_per_m:0,output_per_m:0},openrouter:{input_per_m:.05,output_per_m:.4}},tags:["chat","code","reasoning","open-weight"],huggingface_id:"Qwen/Qwen3-8B",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/Qwen/Qwen3-8B",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.187Z",license:"apache-2.0"},{id:"qwen/qwen-3-coder-8b",name:"Qwen 3 Coder 8B",provider:"Alibaba",family:"Qwen 3 Coder",version:"3",release_date:"2025-06-01",architecture:{type:"dense",total_params_b:8.2,active_params_b:8.2,num_layers:36,hidden_dim:4096,num_heads:32,num_kv_heads:8,head_dim:128,intermediate_dim:12288,vocab_size:152064},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:16.4,weights_fp8_gb:8.2,weights_int4_gb:4.1,kv_per_token_bytes:147456,activation_estimate_gb:.5},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","ollama"],api_pricing:{alibaba:{input_per_m:.15,output_per_m:.15},openrouter:{input_per_m:.22,output_per_m:1.7999999999999998}},tags:["code","chat","open-weight"],release_date_source:"https://qwenlm.github.io/blog/qwen3-coder/",license:"apache-2.0",release_date_verified_at:"2026-06-25T16:20:15.669Z",last_verified_at:"2026-06-25T16:20:15.669Z"},{id:"qwen/qwen-agentworld-35b-a3b",name:"Qwen-AgentWorld-35B-A3B",tags:["transformers","safetensors","qwen3_5_moe","image-text-to-text","qwen","world-model","agent","environment-simulation"],family:"Qwen",memory:{weights_fp8_gb:34.7,weights_bf16_gb:69.4,weights_int4_gb:17.35,kv_per_token_bytes:81920,activation_estimate_gb:32768e-8},version:"1.0",provider:"Qwen",is_partial:!1,api_pricing:{},architecture:{type:"moe",head_dim:256,num_heads:16,hidden_dim:2048,num_layers:40,vocab_size:248320,num_experts:256,num_kv_heads:2,total_params_b:34.7,active_params_b:7,intermediate_dim:512,experts_per_token:8},capabilities:{code:!1,math:!1,tools:!0,vision:!0,reasoning:!1,multilingual:!1,structured_output:!1},release_date:"2026-01-01",context_window:262144,huggingface_id:"Qwen

1/Qwen-AgentWorld-35B-A3B",default_precision:"bf16",supported_frameworks:["vllm"],supported_precisions:["bf16","fp8","int4"],status:"active"},{id:"qwen/qwen3.6-27b",name:"Qwen3.6 27B",family:"Qwen",provider:"Alibaba",is_partial:!0,release_date:"2026-04-27",release_date_source:"https://openrouter.ai/models/qwen/qwen3.6-27b",release_date_verified_at:"2026-06-20T16:00:00Z",last_verified_at:"2026-06-20T16:00:00Z",context_window:262144,deployment:"api",api_pricing:{alibaba:{input_per_m:.6,output_per_m:3.6}},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},description:"Qwen3.6 27B is a mid-size Qwen3.6 model — multimodal, tool-using and reasoning-capable with a 256K-token context window, available through Alibaba Cloud Model Studio."},{id:"qwen/qwen3.6-35b-a3b",name:"Qwen3.6 35B A3B",family:"Qwen",provider:"Alibaba",is_partial:!0,release_date:"2026-04-27",release_date_source:"https://openrouter.ai/models/qwen/qwen3.6-35b-a3b",release_date_verified_at:"2026-06-20T16:00:00Z",last_verified_at:"2026-06-20T16:00:00Z",context_window:262144,deployment:"api",api_pricing:{alibaba:{input_per_m:.375,output_per_m:2.25}},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},description:"Qwen3.6 35B-A3B is a sparse mixture-of-experts Qwen3.6 model (35B total, ~3B active) — multimodal and reasoning-capable with a 256K-token context window, served through Alibaba Cloud Model Studio."},{id:"qwen/qwen3.6-flash",name:"Qwen3.6 Flash",family:"Qwen",provider:"Alibaba",is_partial:!0,release_date:"2026-04-27",release_date_source:"https://openrouter.ai/models/qwen/qwen3.6-flash",release_date_verified_at:"2026-06-20T16:00:00Z",last_verified_at:"2026-06-20T16:00:00Z",context_window:1e6,deployment:"api",api_pricing:{alibaba:{input_per_m:.25,output_per_m:1.5}},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},description:"Qwen3.6 Flash is Alibaba's fast, low-cost Qwen3.6 tier — multimodal, tool-using and reasoning-capable with a 1M-token context window, served through Alibaba Cloud Model Studio."},{id:"qwen/qwen3.6-max-preview",name:"Qwen3.6 Max Preview",family:"Qwen",provider:"Alibaba",is_partial:!0,release_date:"2026-04-27",release_date_source:"https://openrouter.ai/models/qwen/qwen3.6-max-preview",release_date_verified_at:"2026-06-20T16:00:00Z",last_verified_at:"2026-06-20T16:00:00Z",context_window:262144,deployment:"api",api_pricing:{alibaba:{input_per_m:1.3,output_per_m:7.8}},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},description:"Qwen3.6 Max (preview) is Alibaba's high-capability Qwen3.6 tier — a tool-using, reasoning model with a 256K-token context window, served through Alibaba Cloud Model Studio."},{id:"qwen/qwen3.7-max",name:"Qwen3.7 Max",family:"Qwen",provider:"Alibaba",is_partial:!0,release_date:"2026-05-21",release_date_source:"https://openrouter.ai/models/qwen/qwen3.7-max",release_date_verified_at:"2026-06-20T16:00:00Z",last_verified_at:"2026-06-20T16:00:00Z",context_window:1e6,deployment:"api",api_pricing:{alibaba:{input_per_m:2.5,output_per_m:7.5}},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},description:"Qwen3.7 Max is Alibaba's flagship Qwen model — a reasoning and tool-using model with a 1M-token context window, served through Alibaba Cloud Model Studio (DashScope)."},{id:"qwen/qwen3.7-plus",name:"Qwen3.7 Plus",family:"Qwen",provider:"Alibaba",is_partial:!0,release_date:"2026-06-03",release_date_source:"https://openrouter.ai/models/qwen/qwen3.7-plus",release_date_verified_at:"2026-06-20T16:00:00Z",last_verified_at:"2026-06-20T16:00:00Z",context_window:1e6,deployment:"api",api_pricing:{alibaba:{input_per_m:.4,output_per_m:1.6}},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},description:"Qwen3.7 Plus is Alibaba's cost-efficient Qwen tier — multimodal, tool-using and reasoning-capable with a 1M-token context window at roughly one-sixth the price of Qwen3.7 Max."},{id:"qwen/qwen3-asr-0.6b-hf",name:"Qwen3-ASR-0.6B-hf",tags:["transformers","safetensors","qwen3_asr","automatic-speech-recognition","zh","en","yue","ar"],family:"Qwen",memory:{weights_fp8_gb:.8,weights_bf16_gb:1.6,weights_int4_gb:.4,kv_per_token_bytes:114688,activation_estimate_gb:114688e-9},version:"0.6",provider:"Qwen",is_partial:!1,api_pricing:{},architecture:{type:"dense",head_dim:128,num_heads:16,hidden_dim:1024,num_layers:28,vocab_size:151936,num_kv_heads:8,total_params_b:.8,active_params_b:.8,intermediate_dim:3072},capabilities:{code:!1,math:!1,tools:!1,vision:!1,reasoning:!1,multilingual:!1,structured_output:!1},release_date:"2026-01-01",context_window:65536,huggingface_id:"Qwen

1/Qwen3-ASR-0.6B-hf",default_precision:"bf16",supported_frameworks:["vllm"],supported_precisions:["bf16","fp8","int4"],status:"active"},{id:"qwen/qwen3-asr-1.7b-hf",name:"Qwen3-ASR-1.7B-hf",gated:!1,family:"Qwen",version:"1.7",provider:"Qwen",library_name:"transformers",release_date:"2026-01-01",huggingface_id:"Qwen/Qwen3-ASR-1.7B-hf",default_precision:"bf16",release_date_source:"https://huggingface.co/Qwen/Qwen3-ASR-1.7B-hf",supported_precisions:["bf16","fp8","int4"],release_date_verified_at:"2026-07-10T04:34:12.150Z",last_verified_at:"2026-07-10T04:34:12.150Z",modality:"speech-asr",specialized_architecture:{type:"encoder-decoder",total_params_b:2.04,notes:"Qwen3-ASR automatic-speech-recognition model (Qwen3ASRForConditionalGeneration), 2.04B parameters. Encoder-decoder transformer mapping audio input to text transcription. Released open-weight on HuggingFace via the transformers library."},specialized_memory:{weights_bf16_gb:4.08,activation_estimate_gb:1.2},specialized_pricing:{kind:"free_open_weights"},description:"Qwen3-ASR 1.7B is Alibaba's open-weight automatic-speech-recognition model — a 2.04B-parameter encoder-decoder transformer that transcribes audio to text, runnable locally via HuggingFace transformers.",tags:["speech-recognition","asr","audio","open-weight","transformers"]},{id:"replit/replit-code-v1.5-3b",name:"Replit Code v1.5 3B",provider:"Replit",family:"Replit Code",version:"1.5",release_date:"2023-07-18",architecture:{type:"dense",total_params_b:3.3,active_params_b:3.3,num_layers:32,hidden_dim:2560,num_heads:32,num_kv_heads:32,head_dim:80,intermediate_dim:10240,vocab_size:32768},context_window:4096,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:6.6,weights_fp8_gb:3.3,weights_int4_gb:1.65,kv_per_token_bytes:163840,activation_estimate_gb:.4},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","tgi"],api_pricing:{},tags:["code","code-specialized","open-weight","lightweight"],huggingface_id:"replit/replit-code-v1-3b",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/replit/replit-code-v1-3b",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.187Z",license:"cc-by-sa-4.0"},{id:"rinna/nekomata-14b",name:"Nekomata 14B",provider:"Rinna",family:"Nekomata",version:"1",release_date:"2024-01-16",architecture:{type:"dense",total_params_b:14,active_params_b:14,num_layers:40,hidden_dim:5120,num_heads:40,num_kv_heads:8,head_dim:128,intermediate_dim:13824,vocab_size:65536},context_window:4096,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:28,weights_fp8_gb:14,weights_int4_gb:7,kv_per_token_bytes:163840,activation_estimate_gb:1},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","tgi"],api_pricing:{},tags:["japanese","multilingual","open-weight"],huggingface_id:"rinna/nekomata-14b",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/rinna/nekomata-14b",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.187Z",license:"other"},{id:"rwkv/rwkv-6-14b",name:"RWKV-6 14B",provider:"RWKV Foundation",family:"RWKV",version:"6",release_date:"2024-06-01",architecture:{type:"hybrid",total_params_b:14.1,active_params_b:14.1,num_layers:61,hidden_dim:5120,num_heads:1,num_kv_heads:1,head_dim:5120,intermediate_dim:13824,vocab_size:65536},context_window:32768,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:28.2,weights_fp8_gb:14.1,weights_int4_gb:7.05,kv_per_token_bytes:0,activation_estimate_gb:1},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["vllm"],api_pricing:{rwkv:{input_per_m:.2,output_per_m:.2}},tags:["chat","rnn","linear-attention","open-weight"],huggingface_id:"RWKV/v6-Finch-14B-HF",huggingface_id_status:"manual",release_date_source:"https://huggingface.co/RWKV/v6-Finch-14B-HF",last_verified_at:"2026-06-25T16:19:04.534Z",license:"apache-2.0"},{id:"sail/infoxlm-large",name:"InfoXLM Large",provider:"Microsoft SAIL",family:"InfoXLM",version:"1",release_date:"2021-07-15",architecture:{type:"dense",total_params_b:.55,active_params_b:.55,num_layers:24,hidden_dim:1024,num_heads:16,num_kv_heads:16,head_dim:64,intermediate_dim:4096,vocab_size:250002},context_window:512,supported_precisions:["bf16","fp8"],default_precision:"bf16",memory:{weights_bf16_gb:1.1,weights_fp8_gb:.55,weights_int4_gb:.275,kv_per_token_bytes:98304,activation_estimate_gb:.1},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["tgi"],api_pricing:{},tags:["embedding","multilingual","cross-lingual","open-weight"],huggingface_id:"microsoft/infoxlm-large",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/microsoft/infoxlm-large",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-16T09:29:53.000Z"},{id:"salesforce/codegen2-16b",name:"CodeGen2 16B",provider:"Salesforce",family:"CodeGen2",version:"2",release_date:"2023-05-10",architecture:{type:"dense",total_params_b:16,active_params_b:16,num_layers:34,hidden_dim:6144,num_heads:48,num_kv_heads:48,head_dim:128,intermediate_dim:24576,vocab_size:51200},context_window:2048,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:32,weights_fp8_gb:16,weights_int4_gb:8,kv_per_token_bytes:417792,activation_estimate_gb:1.2},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","tgi"],api_pricing:{},tags:["code","code-specialized","open-weight"],huggingface_id:"Salesforce/codegen2-16B_P",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/Salesforce/codegen2-16B_P",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.187Z",license:"apache-2.0"},{id:"sberbank/gigachat-20b",name:"GigaChat 20B",provider:"Sberbank",family:"GigaChat",version:"1",release_date:"2023-11-24",architecture:{type:"dense",total_params_b:20,active_params_b:20,num_layers:44,hidden_dim:6144,num_heads:48,num_kv_heads:8,head_dim:128,intermediate_dim:16384,vocab_size:128e3},context_window:8192,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:40,weights_fp8_gb:20,weights_int4_gb:10,kv_per_token_bytes:180224,activation_estimate_gb:1.2},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","tgi"],api_pricing:{},tags:["chat","multilingual","russian","open-weight"],huggingface_id:"ai-sage/GigaChat-20B-A3B-instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/ai-sage/GigaChat-20B-A3B-instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.187Z",license:"mit"},{id:"scientific/sciglm-6b",name:"SciGLM 6B",provider:"Tsinghua",family:"SciGLM",version:"1",release_date:"2024-01-15",architecture:{type:"dense",total_params_b:6.2,active_params_b:6.2,num_layers:28,hidden_dim:4096,num_heads:32,num_kv_heads:2,head_dim:128,intermediate_dim:13696,vocab_size:65024},context_window:8192,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:12.4,weights_fp8_gb:6.2,weights_int4_gb:3.1,kv_per_token_bytes:14336,activation_estimate_gb:.6},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!0,multilingual:!1,reasoning:!0},supported_frameworks:["vllm","tgi"],api_pricing:{},tags:["scientific","domain-specific","open-weight"],huggingface_id:"zd21/SciGLM-6B",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/zd21/SciGLM-6B",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.187Z",license:"apache-2.0"},{id:"sensetime/internlm-20b",name:"InternLM 20B",provider:"SenseTime",family:"InternLM",version:"1",release_date:"2023-09-20",architecture:{type:"dense",total_params_b:20,active_params_b:20,num_layers:60,hidden_dim:5120,num_heads:40,num_kv_heads:40,head_dim:128,intermediate_dim:13824,vocab_size:103168},context_window:16384,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:40,weights_fp8_gb:20,weights_int4_gb:10,kv_per_token_bytes:614400,activation_estimate_gb:1.5},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","tgi"],api_pricing:{},tags:["chinese","multilingual","open-weight"],huggingface_id:"internlm/internlm-20b",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/internlm/internlm-20b",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.188Z",license:"apache-2.0"},{id:"sentence-transformers/all-minilm-l6-v2",name:"All MiniLM L6 v2",provider:"Sentence Transformers",family:"MiniLM",version:"2",release_date:"2021-08-05",architecture:{type:"dense",total_params_b:.0227,active_params_b:.0227,num_layers:6,hidden_dim:384,num_heads:12,num_kv_heads:12,head_dim:32,intermediate_dim:1536,vocab_size:30522},context_window:256,supported_precisions:["bf16","fp8"],default_precision:"bf16",memory:{weights_bf16_gb:.0454,weights_fp8_gb:.0227,weights_int4_gb:.01135,kv_per_token_bytes:9216,activation_estimate_gb:.01},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["tgi","ollama"],api_pricing:{huggingface:{input_per_m:.005,output_per_m:.005}}

1,tags:["embedding","lightweight","open-weight"],huggingface_id:"Xenova/all-MiniLM-L6-v2",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/Xenova/all-MiniLM-L6-v2",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.188Z",license:"apache-2.0"},{id:"sesame/csm-1b",name:"CSM-1B",provider:"Sesame",family:"CSM",version:"1",release_date:"2025-03-05",architecture:{type:"dense",total_params_b:1,active_params_b:1,num_layers:12,hidden_dim:1024,num_heads:16,num_kv_heads:16,head_dim:64,intermediate_dim:4096,vocab_size:5e4},context_window:4096,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:2,weights_fp8_gb:1,weights_int4_gb:.5,kv_per_token_bytes:24576,activation_estimate_gb:.1},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["ollama"],api_pricing:{},tags:["speech","conversational-ai","open-weight"],huggingface_id:"audo/csm-1b",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/audo/csm-1b",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-16T09:29:53.000Z"},{id:"snowflake/arctic-128x3b",name:"Snowflake Arctic 128x3B",provider:"Snowflake",family:"Arctic",version:"1-128x3b",release_date:"2024-04-24",architecture:{type:"moe",total_params_b:395,active_params_b:17,num_layers:35,hidden_dim:7168,num_heads:56,num_kv_heads:8,head_dim:128,intermediate_dim:4864,vocab_size:32e3,num_experts:128,experts_per_token:2},context_window:4096,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:790,weights_fp8_gb:395,weights_int4_gb:197.5,kv_per_token_bytes:143360,activation_estimate_gb:2},capabilities:{tools:!1,vision:!1,structured_output:!0,code:!0,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang"],api_pricing:{},tags:["chat","moe","sparse","open-weight","enterprise","estimated-architecture"],release_date_source:"https://www.snowflake.com/en/blog/arctic-open-efficient-foundation-language-models-snowflake/",license:"apache-2.0",release_date_verified_at:"2026-06-25T16:20:15.669Z",last_verified_at:"2026-06-25T16:20:15.669Z"},{id:"snowflake/arctic-480b",name:"Snowflake Arctic 480B",provider:"Snowflake",family:"Arctic",version:"1",release_date:"2024-04-24",architecture:{type:"moe",total_params_b:480,active_params_b:17,num_layers:35,hidden_dim:7168,num_heads:56,num_kv_heads:8,head_dim:128,intermediate_dim:4864,vocab_size:32e3,num_experts:128,experts_per_token:2},context_window:4096,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:960,weights_fp8_gb:480,weights_int4_gb:240,kv_per_token_bytes:143360,activation_estimate_gb:2},capabilities:{tools:!1,vision:!1,structured_output:!0,code:!0,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang"],api_pricing:{snowflake:{input_per_m:1.5,output_per_m:1.5}},tags:["chat","moe","sparse","open-weight","enterprise"],huggingface_id:"Snowflake/snowflake-arctic-instruct",huggingface_id_status:"manual",release_date_source:"https://huggingface.co/Snowflake/snowflake-arctic-instruct",last_verified_at:"2026-06-25T16:19:04.534Z",license:"apache-2.0"},{id:"stabilityai/sdxl-1.0",name:"Stable Diffusion XL 1.0",provider:"Stability AI",family:"Stable Diffusion",version:"xl-1.0",release_date:"2023-07-26",architecture:{type:"dense",total_params_b:3.5,active_params_b:3.5,num_layers:32,hidden_dim:2560,num_heads:32,num_kv_heads:32,head_dim:80,intermediate_dim:10240,vocab_size:49408},context_window:77,supported_precisions:["bf16","fp8"],default_precision:"bf16",memory:{weights_bf16_gb:7,weights_fp8_gb:3.5,weights_int4_gb:1.75,kv_per_token_bytes:163840,activation_estimate_gb:2},capabilities:{tools:!1,vision:!0,structured_output:!1,code:!1,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:[],api_pricing:{},tags:["image-generation","open-weight"],huggingface_id:"stabilityai/stable-diffusion-xl-base-1.0",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/stabilityai/stable-diffusion-xl-base-1.0",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.188Z",license:"openrail++"},{id:"stabilityai/stablelm-2-12b",name:"StableLM 2 12B",provider:"Stability AI",family:"StableLM 2",version:"2",release_date:"2024-01-19",architecture:{type:"dense",total_params_b:12.1,active_params_b:12.1,num_layers:40,hidden_dim:5120,num_heads:32,num_kv_heads:8,head_dim:160,intermediate_dim:13824,vocab_size:100352},context_window:4096,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:24.2,weights_fp8_gb:12.1,weights_int4_gb:6.05,kv_per_token_bytes:204800,activation_estimate_gb:1},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","ollama"],api_pricing:{stabilityai:{input_per_m:.25,output_per_m:.25}},tags:["chat","open-weight"],huggingface_id:"stabilityai/stablelm-2-12b",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/stabilityai/stablelm-2-12b",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.188Z",license:"other"},{id:"stabilityai/stablelm-zephyr-3b",name:"StableLM Zephyr 3B",provider:"Stability AI",family:"StableLM",version:"2",release_date:"2023-11-29",architecture:{type:"dense",total_params_b:3,active_params_b:3,num_layers:32,hidden_dim:2560,num_heads:32,num_kv_heads:32,head_dim:80,intermediate_dim:6912,vocab_size:100352},context_window:4096,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:6,weights_fp8_gb:3,weights_int4_gb:1.5,kv_per_token_bytes:327680,activation_estimate_gb:.3},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","ollama"],api_pricing:{},tags:["chat","lightweight","edge","open-weight"],huggingface_id:"stabilityai/stablelm-zephyr-3b",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/stabilityai/stablelm-zephyr-3b",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.188Z",license:"other"},{id:"stepfun-ai/nextstep-1.1-pretrain-256px",name:"NextStep-1.1-Pretrain-256px",family:"stepfun-ai",provider:"stepfun-ai",is_partial:!0,release_date:"2026-02-16",last_verified_at:"2026-06-22T10:34:00.673Z",release_date_source:"https://huggingface.co/stepfun-ai/NextStep-1.1-Pretrain-256px",release_date_verified_at:"2026-06-22T10:34:00.673Z",status:"draft"},{id:"stepfun-ai/step-3.5-flash-base-midtrain",name:"Step-3.5-Flash-Base-Midtrain",family:"stepfun-ai",license:"apache-2.0",version:"3.5",provider:"stepfun-ai",is_partial:!0,capabilities:{code:!0,math:!1,tools:!1,vision:!1,reasoning:!1,multilingual:!1,structured_output:!1},release_date:"2026-03-02",last_verified_at:"2026-06-22T11:43:20.828Z",release_date_source:"https://huggingface.co/stepfun-ai/Step-3.5-Flash-Base-Midtrain",release_date_verified_at:"2026-06-22T11:43:20.828Z",status:"draft"},{id:"stepfun-ai/step-3.7-flash-nvfp4",name:"Step-3.7-Flash-NVFP4",family:"stepfun-ai",provider:"stepfun-ai",is_partial:!0,release_date:"2026-05-27",last_verified_at:"2026-06-22T10:30:53.992Z",release_date_source:"https://huggingface.co/stepfun-ai/Step-3.7-Flash-NVFP4",release_date_verified_at:"2026-06-22T10:30:53.992Z",status:"draft"},{id:"teknium/openhermes-2.5-7b",name:"OpenHermes 2.5 7B",provider:"Teknium",family:"OpenHermes",version:"2.5",release_date:"2023-10-29",architecture:{type:"dense",total_params_b:7,active_params_b:7,num_layers:32,hidden_dim:4096,num_heads:32,num_kv_heads:8,head_dim:128,intermediate_dim:14336,vocab_size:32e3},context_window:32768,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:14,weights_fp8_gb:7,weights_int4_gb:3.5,kv_per_token_bytes:131072,activation_estimate_gb:1},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","ollama"],api_pricing:{together:{input_per_m:.2,output_per_m:.2}},tags:["chat","code","instruction-following","open-weight"],huggingface_id:"teknium/OpenHermes-2.5-Mistral-7B",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/teknium/OpenHermes-2.5-Mistral-7B",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.188Z",license:"apache-2.0"},{id:"tencent/hy-embodied-0.5-vla-robotwin",name:"Hy-Embodied-0.5-VLA-RoboTwin",tags:["safetensors","robotics","VLA","bimanual","RoboTwin","imitation-learning","fine-tuned"],family:"tencent",memory:{weights_fp8_gb:4.5,weights_bf16_gb:9,weights_int4_gb:2.25,kv_per_token_bytes:131072,activation_estimate_gb:524288e-9},version:"0.5",provider:"tencent",api_pricing:{},architecture:{type:"dense",head_dim:128,num_heads:32,hidden_dim:4096,num_layers:32,vocab_size:32e3,num_kv_heads:8,total_params_b:4.5,active_params_b:4.5,intermediate_dim:11008},capabilities:{code:!1,math:!1,tools:!1,vision:!1,reasoning:!1,multilingual:!1,structured_output:!1},release_date:"2026-06-11",context_window:8192,huggingface_id:"tencent/Hy-Embodied-0.5-VLA-RoboTwin",default_precision:"bf16",supp

1orted_frameworks:["vllm"],supported_precisions:["bf16"],status:"draft",license:"apache-2.0",last_verified_at:"2026-06-25T15:18:14.188Z"},{id:"tencent/hy-embodied-0.5-vla-umi",name:"Hy-Embodied-0.5-VLA-UMI",tags:["safetensors","robotics","VLA","bimanual","UMI","imitation-learning","pre-trained"],family:"tencent",memory:{weights_fp8_gb:4.5,weights_bf16_gb:9,weights_int4_gb:2.25,kv_per_token_bytes:131072,activation_estimate_gb:524288e-9},version:"0.5",provider:"tencent",api_pricing:{},architecture:{type:"dense",head_dim:128,num_heads:32,hidden_dim:4096,num_layers:32,vocab_size:32e3,num_kv_heads:8,total_params_b:4.5,active_params_b:4.5,intermediate_dim:11008},capabilities:{code:!1,math:!1,tools:!1,vision:!1,reasoning:!1,multilingual:!1,structured_output:!1},release_date:"2026-06-11",context_window:8192,huggingface_id:"tencent/Hy-Embodied-0.5-VLA-UMI",default_precision:"bf16",supported_frameworks:["vllm"],supported_precisions:["bf16"],status:"draft",license:"apache-2.0",last_verified_at:"2026-06-25T15:18:14.188Z"},{id:"tencent/universal_audio_tokenizer",name:"Universal_Audio_Tokenizer",family:"tencent",license:"other",provider:"tencent",release_date:"2026-05-27",release_date_source:"https://huggingface.co/tencent/Universal_Audio_Tokenizer",huggingface_id:"tencent/Universal_Audio_Tokenizer",release_date_verified_at:"2026-06-23T11:31:35.151Z",last_verified_at:"2026-06-23T11:31:35.151Z",modality:"speech-asr",specialized_architecture:{type:"encoder",notes:"Compact audio tokenizer that unifies general audio perception with linguistic alignment for Audio-LLMs. Components: a WhisperVQ encoder plus an AudioDecoder with flow + HIFT modules, trained with Semantic-Acoustic Primitives supervision and a Semantic-Acoustic Equilibrium mechanism. 25 Hz frame rate, 8192-entry codebook, ~325 bits/sec. Parameter count not published."},specialized_pricing:{kind:"free_open_weights"},description:"Tencent's Universal Audio Tokenizer is an open-weight audio tokenizer that bridges acoustic and semantic understanding for Audio-LLMs. It encodes speech, sound and music into discrete tokens (25 Hz, 8192 codebook, ~325 bps) for LLM integration, TTS and audio perception.",tags:["audio","tokenizer","audio-llm","speech","open-weight"]},{id:"thudm/chatglm3-6b",name:"ChatGLM3 6B",provider:"Tsinghua University",family:"ChatGLM3",version:"3",release_date:"2023-10-27",architecture:{type:"dense",total_params_b:6,active_params_b:6,num_layers:28,hidden_dim:4096,num_heads:32,num_kv_heads:2,head_dim:128,intermediate_dim:13696,vocab_size:65024},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:12,weights_fp8_gb:6,weights_int4_gb:3,kv_per_token_bytes:28672,activation_estimate_gb:.8},capabilities:{tools:!0,vision:!1,structured_output:!1,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","ollama"],api_pricing:{},tags:["chat","code","open-weight","multilingual"],huggingface_id:"zai-org/chatglm3-6b",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/zai-org/chatglm3-6b",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-16T09:29:53.000Z"},{id:"thudm/cogvlm2-19b",name:"CogVLM2 19B",provider:"THUDM",family:"CogVLM2",version:"2",release_date:"2024-05-24",architecture:{type:"dense",total_params_b:19,active_params_b:19,num_layers:40,hidden_dim:4096,num_heads:32,num_kv_heads:32,head_dim:128,intermediate_dim:14336,vocab_size:128256},context_window:8192,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:38,weights_fp8_gb:19,weights_int4_gb:9.5,kv_per_token_bytes:655360,activation_estimate_gb:1.2},capabilities:{tools:!1,vision:!0,structured_output:!1,code:!1,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi"],api_pricing:{},tags:["chat","vision","vision-language","open-weight"],huggingface_id:"zai-org/cogvlm2-llama3-chat-19B",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/zai-org/cogvlm2-llama3-chat-19B",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.188Z",license:"other"},{id:"tiiuae/falcon-11b",name:"Falcon 11B",provider:"TII UAE",family:"Falcon",version:"2",release_date:"2024-05-13",architecture:{type:"dense",total_params_b:11,active_params_b:11,num_layers:60,hidden_dim:4096,num_heads:32,num_kv_heads:8,head_dim:128,intermediate_dim:14336,vocab_size:65024},context_window:8192,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:22,weights_fp8_gb:11,weights_int4_gb:5.5,kv_per_token_bytes:245760,activation_estimate_gb:.8},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi"],api_pricing:{},tags:["chat","open-weight"],huggingface_id:"tiiuae/falcon-11B",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/tiiuae/falcon-11B",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.188Z",license:"unknown"},{id:"tiiuae/falcon-3-10b",name:"Falcon 3 10B",provider:"TII UAE",family:"Falcon",version:"3",release_date:"2024-12-12",architecture:{type:"dense",total_params_b:10.3,active_params_b:10.3,num_layers:40,hidden_dim:4096,num_heads:32,num_kv_heads:8,head_dim:128,intermediate_dim:14336,vocab_size:131072},context_window:32768,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:20.6,weights_fp8_gb:10.3,weights_int4_gb:5.15,kv_per_token_bytes:81920,activation_estimate_gb:.7},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","llama-cpp","ollama"],api_pricing:{},tags:["chat","code","open-weight"],huggingface_id:"tiiuae/Falcon3-10B-Instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/tiiuae/Falcon3-10B-Instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.188Z",license:"other"},{id:"tiiuae/falcon-3-1b",name:"Falcon 3 1B",provider:"TII UAE",family:"Falcon",version:"3",release_date:"2024-12-12",architecture:{type:"dense",total_params_b:1,active_params_b:1,num_layers:18,hidden_dim:2048,num_heads:16,num_kv_heads:4,head_dim:128,intermediate_dim:8192,vocab_size:131072},context_window:8192,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:2,weights_fp8_gb:1,weights_int4_gb:.5,kv_per_token_bytes:18432,activation_estimate_gb:.2},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","llama-cpp","ollama"],api_pricing:{},tags:["chat","open-weight","small"],huggingface_id:"tiiuae/falcon-rw-1b",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/tiiuae/falcon-rw-1b",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.188Z",license:"apache-2.0"},{id:"tiiuae/falcon-3-3b",name:"Falcon 3 3B",provider:"TII UAE",family:"Falcon",version:"3",release_date:"2024-12-12",architecture:{type:"dense",total_params_b:3,active_params_b:3,num_layers:24,hidden_dim:2560,num_heads:20,num_kv_heads:4,head_dim:128,intermediate_dim:10240,vocab_size:131072},context_window:8192,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:6,weights_fp8_gb:3,weights_int4_gb:1.5,kv_per_token_bytes:24576,activation_estimate_gb:.3},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","llama-cpp","ollama"],api_pricing:{},tags:["chat","open-weight","small"],huggingface_id:"tiiuae/Falcon3-3B-Instruct-1.58bit",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/tiiuae/Falcon3-3B-Instruct-1.58bit",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.188Z",license:"other"},{id:"tiiuae/falcon-3-7b",name:"Falcon 3 7B",provider:"TII UAE",family:"Falcon",version:"3",release_date:"2024-12-12",architecture:{type:"dense",total_params_b:7.5,active_params_b:7.5,num_layers:32,hidden_dim:4096,num_heads:32,num_kv_heads:8,head_dim:128,intermediate_dim:14336,vocab_size:131072},context_window:32768,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:15,weights_fp8_gb:7.5,weights_int4_gb:3.75,kv_per_token_bytes:65536,activation_estimate_gb:.5},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","llama-cpp","ollama"],api_pricing:{},tags:["chat","code","open-weight"],huggingface_id:"tiiuae/Falcon3-7B-Instruct",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/tiiuae/Falcon3-7B-Instruct",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.188Z",license:"other"},{id:"tiiuae/falcon-mamba-7b",name:"Falcon Mamba 7B",provider:"TII",family:"Falcon Mamba",version:"1",release_date:"2024-08-12",architecture:{type:"hybrid",total_params_b:7.27,active_params_b:7.27,num_layers:64,hidden_dim:4096,num_heads:1,num_kv_heads:1,head_dim:4096,intermediate_dim:8192,vocab_size:65024},context_window:8192,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:14.54,weights_fp8_gb:7.27,weights_int4_gb:3.635,kv_per_token_bytes:0,activation_estimate_gb:1},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang"],api_pricing:{tii:{input_per_m:.15,output_per_m:.15}},tags:["chat","ssm","mamba","open-weight"],huggingface_id:"tiiuae/falcon-mamba-7b",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/tiiuae/falcon-mamba-7b",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.188Z",license:"other"},{id:"tinyllama/tinyllama-1.1b-chat",name:"TinyLlama 1.1B Chat",provider:"TinyLlama",family:"TinyLlama",version:"1.0-chat",release_date:"2024-01-25",architecture:{type:"dense",total_params_b:1.1,active_params_b:1.1,num_layers:22,hidden_dim:2048,num_heads:32,num_kv_heads:4,head_dim:64,intermediate_dim:5632,vocab_size:32e3},context_window:2048,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:2.2,weights_fp8_gb:1.1,weights_int4_gb:.55,kv_per_token_bytes:11264,activation_estimate_gb:.15},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","ollama","llama-cpp"],api_pricing:{},tags:["chat","open-weight","lightweight","edge"],huggingface_id:"TinyLlama/TinyLlama-1.1B-Chat-v1.0",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/TinyLlama/TinyLlama-1.1B-Chat-v1.0",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.188Z",license:"apache-2.0"},{id:"tinyllama/tinyllama-1.1b",name:"TinyLlama 1.1B",provider:"TinyLlama",family:"TinyLlama",version:"1",release_date:"2024-01-25",architecture:{type:"dense",total_params_b:1.1,active_params_b:1.1,num_layers:22,hidden_dim:2048,num_heads:32,num_kv_heads:4,head_dim:64,intermediate_dim:5632,vocab_size:32e3},context_window:2048,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:2.2,weights_fp8_gb:1.1,weights_int4_gb:.55,kv_per_token_bytes:45056,activation_estimate_gb:.15},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","ollama"],api_pricing:{},tags:["chat","lightweight","edge","open-weight"],huggingface_id:"TinyLlama/TinyLlama-1.1B-Chat-v1.0",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/TinyLlama/TinyLlama-1.1B-Chat-v1.0",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.188Z",license:"apache-2.0"},{id:"together/llama-3.1-70b-turbo",name:"Llama 3.1 70B Turbo",provider:"Together AI",family:"Llama 3.1",version:"3.1-turbo",release_date:"2024-0

17-23",architecture:{type:"dense",total_params_b:70.6,active_params_b:70.6,num_layers:80,hidden_dim:8192,num_heads:64,num_kv_heads:8,head_dim:128,intermediate_dim:28672,vocab_size:128256},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"fp8",memory:{weights_bf16_gb:141.2,weights_fp8_gb:70.6,weights_int4_gb:35.3,kv_per_token_bytes:327680,activation_estimate_gb:2.5},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tensorrt-llm"],api_pricing:{together:{input_per_m:.88,output_per_m:.88},openrouter:{input_per_m:.4,output_per_m:.4}},tags:["chat","code","optimized","open-weight"],release_date_source:"https://docs.together.ai/docs/serverless-models",license:"llama-3.1-community",release_date_verified_at:"2026-06-25T16:20:15.669Z",last_verified_at:"2026-06-25T16:20:15.669Z"},{id:"upstage/solar-10.7b",name:"SOLAR 10.7B",provider:"Upstage",family:"SOLAR",version:"1",release_date:"2023-12-13",architecture:{type:"dense",total_params_b:10.7,active_params_b:10.7,num_layers:48,hidden_dim:4096,num_heads:32,num_kv_heads:8,head_dim:128,intermediate_dim:11008,vocab_size:32e3},context_window:4096,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:21.4,weights_fp8_gb:10.7,weights_int4_gb:5.35,kv_per_token_bytes:196608,activation_estimate_gb:1},capabilities:{tools:!1,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","ollama"],api_pricing:{upstage:{input_per_m:.3,output_per_m:.3}},tags:["chat","open-weight","depth-upscaled"],huggingface_id:"upstage/SOLAR-10.7B-Instruct-v1.0",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/upstage/SOLAR-10.7B-Instruct-v1.0",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.188Z",license:"cc-by-nc-4.0"},{id:"upstage/solar-pro-22b",name:"Solar Pro 22B",provider:"Upstage",family:"Solar",version:"pro",release_date:"2024-09-05",architecture:{type:"dense",total_params_b:22,active_params_b:22,num_layers:48,hidden_dim:4096,num_heads:32,num_kv_heads:8,head_dim:128,intermediate_dim:14336,vocab_size:32e3},context_window:4096,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:44,weights_fp8_gb:22,weights_int4_gb:11,kv_per_token_bytes:196608,activation_estimate_gb:1.5},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm","ollama"],api_pricing:{upstage:{input_per_m:.5,output_per_m:.5}},tags:["chat","code","instruction-following","open-weight","estimated-architecture"],huggingface_id:"upstage/solar-pro-preview-instruct",huggingface_id_status:"manual",release_date_source:"https://huggingface.co/upstage/solar-pro-preview-instruct",last_verified_at:"2026-06-25T16:19:04.535Z",license:"mit"},{id:"vikhyat/moondream-2b",name:"Moondream 2B",provider:"Vikhyat",family:"Moondream",version:"2",release_date:"2025-02-14",architecture:{type:"dense",total_params_b:1.86,active_params_b:1.86,num_layers:24,hidden_dim:2048,num_heads:16,num_kv_heads:16,head_dim:128,intermediate_dim:5632,vocab_size:51200},context_window:2048,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:3.72,weights_fp8_gb:1.86,weights_int4_gb:.93,kv_per_token_bytes:98304,activation_estimate_gb:.2},capabilities:{tools:!1,vision:!0,structured_output:!1,code:!1,math:!1,multilingual:!1,reasoning:!1},supported_frameworks:["ollama","vllm"],api_pricing:{},tags:["vision","edge","open-weight"],huggingface_id:"vikhyatk/moondream2",huggingface_id_status:"manual",release_date_source:"https://huggingface.co/vikhyatk/moondream2",last_verified_at:"2026-06-25T16:19:04.535Z",license:"apache-2.0"},{id:"wizardlm/wizardcoder-33b",name:"WizardCoder 33B",provider:"WizardLM",family:"WizardCoder",version:"1",release_date:"2024-01-04",architecture:{type:"dense",total_params_b:33,active_params_b:33,num_layers:62,hidden_dim:7168,num_heads:56,num_kv_heads:8,head_dim:128,intermediate_dim:19200,vocab_size:32256},context_window:16384,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:66,weights_fp8_gb:33,weights_int4_gb:16.5,kv_per_token_bytes:253952,activation_estimate_gb:1.8},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!0,math:!0,multilingual:!1,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","tensorrt-llm"],api_pricing:{},tags:["code","code-specialized","open-weight"],huggingface_id:"WizardLMTeam/WizardCoder-33B-V1.1",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/WizardLMTeam/WizardCoder-33B-V1.1",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-16T09:29:53.000Z"},{id:"xai/grok-2",name:"Grok-2",provider:"xAI",family:"Grok",version:"2",release_date:"2024-08-13",architecture:{type:"moe",total_params_b:314,active_params_b:50,num_layers:64,hidden_dim:8192,num_heads:64,num_kv_heads:8,head_dim:128,intermediate_dim:24576,vocab_size:131072,num_experts:8,experts_per_token:2},context_window:131072,supported_precisions:["bf16","fp8"],default_precision:"bf16",memory:{weights_bf16_gb:628,weights_fp8_gb:314,weights_int4_gb:157,kv_per_token_bytes:262144,activation_estimate_gb:3},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm"],api_pricing:{xai:{input_per_m:2,output_per_m:10}},tags:["chat","moe","estimated-architecture","api-only","flagship"],huggingface_id_status:"n_a",release_date_source:"https://x.ai/",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-16T09:29:53.000Z"},{id:"xai/grok-3-mini",name:"Grok-3 Mini",provider:"xAI",family:"Grok",version:"3-mini",release_date:"2025-03-14",architecture:{type:"dense",total_params_b:33,active_params_b:33,num_layers:64,hidden_dim:5120,num_heads:40,num_kv_heads:8,head_dim:128,intermediate_dim:20480,vocab_size:131072},context_window:131072,supported_precisions:["bf16","fp8"],default_precision:"bf16",memory:{weights_bf16_gb:66,weights_fp8_gb:33,weights_int4_gb:16.5,kv_per_token_bytes:262144,activation_estimate_gb:2},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm"],api_pricing:{xai:{input_per_m:.3,output_per_m:.5},openrouter:{input_per_m:.3,output_per_m:.5}},tags:["chat","code","reasoning","proprietary","estimated-architecture","api-only"],huggingface_id_status:"n_a",release_date_source:"https://x.ai/",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-16T09:29:53.000Z"},{id:"xai/grok-3",name:"Grok-3",provider:"xAI",family:"Grok",version:"3",release_date:"2025-02-17",architecture:{type:"dense",total_params_b:314,active_params_b:314,num_layers:96,hidden_dim:12288,num_heads:96,num_kv_heads:16,head_dim:128,intermediate_dim:49152,vocab_size:131072},context_window:131072,supported_precisions:["bf16","fp8"],default_precision:"bf16",memory:{weights_bf16_gb:628,weights_fp8_gb:314,weights_int4_gb:157,kv_per_token_bytes:786432,activation_estimate_gb:5},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm"],api_pricing:{xai:{input_per_m:3,output_per_m:15},openrouter:{input_per_m:3,output_per_m:15}},tags:["chat","code","reasoning","math","proprietary","estimated-architecture","api-only","flagship"],huggingface_id_status:"n_a",release_date_source:"https://x.ai/",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-05-16T09:29:53.000Z"},{name:"Grok 4.3",provider:"xAI",family:"Grok",version:"4.3",release_date:"2026-05-15",release_date_source:"https://docs.x.ai/docs/models",release_date_verified_at:"2026-05-16T14:43:00Z",last_verified_at:"2026-05-16T14:43:00Z",context_window:1e6,supported_precisions:["bf16"],default_precision:"bf16",architecture:{type:"moe",total_params_b:500,active_params_b:90,num_layers:80,hidden_dim:10240,num_heads:80,num_kv_heads:10,head_dim:128,intermediate_dim:36864,vocab_size:2e5,num_experts:16,experts_per_token:2},memory:{weights_bf16_gb:1e3,weights_fp8_gb:500,weights_int4_gb:250,kv_per_token_bytes:204800,activation_estimate_gb:4},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm"],api_pricing:{xai:{input_per_m:1.25,output_per_m:2.5}},is_partial:!1,id:"xai/grok-4-3"},{id:"xai/grok-4",name:"Grok 4",provider:"xAI",family:"Grok",version:"4",release_date:"2025-07-09",release_date_source:"https://openrouter.ai/x-ai/grok-4",release_date_verified_at:"2026-05-16T14:43:00Z",last_verified_at:"2026-05-16T14:43:00Z",is_partial:!1,tags:["frontier"],context_window:256e3,supported_precisions:["bf16"],default_precision:"bf16",architecture:{type:"moe",total_params_b:400,active_params_b:80,num_layers:80,hidden_dim:10240,num_heads:80,num_kv_heads:10,head_dim:128,intermediate_dim:36864,vocab_size:2e5,num_experts:16,experts_per_token:2},memory:{weights_bf16_gb:800,weights_fp8_gb:400,weights_int4_gb:200,kv_per_token_bytes:204800,activation_estimate_gb:4},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm"],api_pricing:{}},{id:"yandex/yalm-100b",name:"YaLM 100B",provider:"Yandex",family:"YaLM",version:"1",release_date:"2022-06-22",architecture:{type:"dense",total_params_b:100,active_params_b:100,num_layers:80,hidden_dim:10240,num_heads:80,num_kv_heads:80,head_dim:128,intermediate_dim:27308,vocab_size:128e3},context_window:2048,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:200,weights_fp8_gb:100,weights_int4_gb:50,kv_per_token_bytes:3276800,activation_estimate_gb:3.5},capabilities:{tools:!1,vision:!1,structured_output:!1,code:!1,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","tgi"],api_pricing:{},tags:["chat","multilingual","russian","open-weight"],huggingface_id:"yandex/yalm-100b",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/yandex/yalm-100b",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.188Z",license:"apache-2.0"},{id:"yilai/yi-6b-200k",name:"Yi 6B 200K",provider:"01.AI",family:"Yi",version:"1-200k",release_date:"2023-11-02",architecture:{type:"dense",total_params_b:6,active_params_b:6,num_layers:32,hidden_dim:4096,num_heads:32,num_kv_heads:4,head_dim:128,intermediate_dim:11008,vocab_size:64e3},context_window:2e5,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:12,weights_fp8_gb:6,weights_int4_gb:3,kv_per_token_bytes:65536,activation_estimate_gb:.8},capabilities:{tools:!1,vision:!1,structured_output:!0,code:!0,math:!1,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","ollama"],api_pricing:{},tags:["chat","long-context","open-weight"],huggingface_id:"01-ai/Yi-6B-200K",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/01-ai/Yi-6B-200K",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.189Z",license:"apache-2.0"},{id:"zai-org/glm-4.7-flash",name:"GLM-4.7-Flash",tags:["transformers","safetensors","glm4_moe_lite","text-generation","conversational","en","zh","eval-results"],family:"GLM-4",memory:{weights_fp8_gb:31.2,weights_bf16_gb:62.4,weights_int4_gb:15.6,kv_per_token_bytes:721920,activation_estimate_gb:385024e-9},version:"4.7",provider:"zai-org",is_partial:!1,api_pricing:{},architecture:{type:"moe",head_dim:192,num_heads:20,hidden_dim:2048,num_layers:47,vocab_size:154880,num_experts:64,num_kv_heads:20,total_params_b:31.2,active_params_b:3,intermediate_dim:10240,experts_per_token:4},capabilities:{code:!0,math:!0,tools:!0,vision:!1,reasoning:!0,multilingual:!0,structured_output:!0},release_date:"2026-01-19",context_window:202752,huggingface_id:"zai-org/GLM-4.7-Flash",default_precision:"bf16",supp

1orted_frameworks:["vllm","sglang"],supported_precisions:["bf16","fp8"],status:"active"},{id:"zai-org/scail-2",name:"SCAIL-2",family:"zai-org",license:"mit",version:"2",provider:"zai-org",release_date:"2026-06-09",huggingface_id:"zai-org/SCAIL-2",last_verified_at:"2026-06-23T11:43:05.427Z",release_date_source:"https://huggingface.co/zai-org/SCAIL-2",release_date_verified_at:"2026-06-23T11:43:05.427Z",modality:"video-fx",specialized_architecture:{type:"diffusion",notes:"Open-source image-to-video model for end-to-end controlled character animation. Uses a Unified Motion Transfer Interface with dedicated masking channels and a RoPE design, integrating a Wan VAE and a T5 text encoder. Removes dependence on intermediate skeleton maps for direct end-to-end driving. Parameter count not published."},specialized_pricing:{kind:"free_open_weights"},description:"SCAIL-2 is an open-source image-to-video model for end-to-end controlled character animation. It drives a source image directly from a reference motion without intermediate skeleton maps, using a Unified Motion Transfer Interface built on a Wan VAE and T5 encoder.",tags:["image-to-video","video-generation","character-animation","diffusion","open-weight"]},{id:"zhipu/chatglm4-9b",name:"ChatGLM4 9B",provider:"Zhipu AI",family:"ChatGLM",version:"4",release_date:"2024-06-05",architecture:{type:"dense",total_params_b:9.4,active_params_b:9.4,num_layers:40,hidden_dim:4096,num_heads:32,num_kv_heads:2,head_dim:128,intermediate_dim:13696,vocab_size:151552},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:18.8,weights_fp8_gb:9.4,weights_int4_gb:4.7,kv_per_token_bytes:20480,activation_estimate_gb:.8},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi"],api_pricing:{},tags:["chinese","multilingual","chat","open-weight"],huggingface_id:"npc0/chatglm-4-9b-int4",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/npc0/chatglm-4-9b-int4",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.189Z",license:"other"},{id:"zhipu/glm-4-9b",name:"GLM-4 9B",provider:"Zhipu AI",family:"GLM-4",version:"4",release_date:"2024-06-06",architecture:{type:"dense",total_params_b:9.4,active_params_b:9.4,num_layers:40,hidden_dim:4096,num_heads:32,num_kv_heads:2,head_dim:128,intermediate_dim:13696,vocab_size:151552},context_window:131072,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:18.8,weights_fp8_gb:9.4,weights_int4_gb:4.7,kv_per_token_bytes:40960,activation_estimate_gb:1},capabilities:{tools:!0,vision:!1,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!1},supported_frameworks:["vllm","sglang","tgi","ollama"],api_pricing:{zhipu:{input_per_m:.15,output_per_m:.15}},tags:["chat","open-weight","instruction-following"],huggingface_id:"zai-org/glm-4-9b-chat",huggingface_id_status:"auto_resolved",release_date_source:"https://huggingface.co/zai-org/glm-4-9b-chat",release_date_verified_at:"2026-05-16T09:29:53.000Z",last_verified_at:"2026-06-25T15:18:14.189Z",license:"other"},{id:"zhipu/glm-5",name:"GLM-5",provider:"Zhipu AI",family:"GLM-5",version:"5",release_date:"2026-02-12",release_date_source:"https://docs.z.ai/release-notes/new-released",release_date_verified_at:"2026-05-13T14:42:10Z",last_verified_at:"2026-06-25T15:18:14.189Z",architecture:{type:"dense",total_params_b:200,active_params_b:200,num_layers:80,hidden_dim:12288,num_heads:96,num_kv_heads:8,head_dim:128,intermediate_dim:40960,vocab_size:152064},context_window:128e3,supported_precisions:["bf16","fp8","int4"],default_precision:"bf16",memory:{weights_bf16_gb:400,weights_fp8_gb:200,weights_int4_gb:100,kv_per_token_bytes:327680,activation_estimate_gb:4},capabilities:{tools:!0,vision:!0,structured_output:!0,code:!0,math:!0,multilingual:!0,reasoning:!0},supported_frameworks:["vllm","sglang","tgi"],api_pricing:{zhipu:{input_per_m:2,output_per_m:6},together:{input_per_m:1,output_per_m:3.2},novita:{input_per_m:1,output_per_m:3.2},fireworks:{input_per_m:0,output_per_m:0},zai:{input_per_m:0,output_per_m:0},featherless:{input_per_m:0,output_per_m:0},openrouter:{input_per_m:.6,output_per_m:1.92}},tags:["chat","code","reasoning","flagship","multilingual","open-weight"],huggingface_id:"zai-org/GLM-5",huggingface_id_status:"auto_resolved",license:"mit"}].map((e,i)=>{let t=u.safeParse(e);if(!t.success){let a=e&&"object"==typeof e&&"id"in e?String(e.id):`<entry #${i}>`;throw Error(`[data/models] schema validation failed for ${a}: ${JSON.stringify(t.error.issues)}`)}return t.data});function h(e){return void 0===e.status||"active"===e.status}let b=c.filter(e=>!0!==e.is_partial).filter(g).filter(h).map(e=>e),f=c.filter(e=>!0!==e.is_partial).filter(e=>!g(e)).filter(h).map(e=>e),v=c.filter(e=>!0===e.is_partial).filter(h).map(e=>e);c.filter(e=>!0!==e.is_partial).filter(g).map(e=>e),c.filter(e=>!0!==e.is_partial).filter(e=>!g(e)).map(e=>e),c.filter(e=>!0===e.is_partial).map(e=>

1e),e.s(["getModel",0,function(e){return b.find(i=>i.id===e)},"models",0,b,"partialModels",0,v,"searchModels",0,function(e){let i=e.toLowerCase();return b.filter(e=>e.name.toLowerCase().includes(i)||e.id.toLowerCase().includes(i)||e.family.toLowerCase().includes(i)||e.provider.toLowerCase().includes(i))},"specializedModels",0,f],583517)}]);