AI Agent StoreRead input and output prices together
A model priced at $2 input and $10 output per million tokens costs $40 for 1,000 calls of 10,000 input and 2,000 output tokens each. Different tokenizers can turn the same text into different token counts.
THE MODEL SELECTION SERIES
LLM API pricing calculator
API prices are usually quoted per million tokens, but your budget depends on how much context you send and how much output the model bills. Change the calculator to compare the same workload across providers.
Snapshot · LiveBench question set: 25 June 2026 · View every source ↗ · Download snapshot ↓
Reasoning, caching and long context change the bill
Include billed reasoning tokens in output. Cache reads may be cheaper while cache writes can cost more. Some providers charge higher rates for large prompts or premium service tiers. Our estimate excludes those adjustments; check the linked source before committing a budget.
Direct API and routed prices are different offers
We use verified direct prices for selected OpenAI and Anthropic models and explicitly labeled OpenRouter catalog rates for other models. A catalog entry can route across endpoints with different prices and limits. Each profile shows the exact catalog identifier and any observed price difference.
Use Practical Value as a preference, not a financial forecast
The default score combines 70% task-weighted capability with 30% affordability rank within the snapshot. Move the slider to express your tradeoff. It does not predict task success, retries, hosting charges or actual monthly savings.
YOUR WORKLOAD. YOUR PRIORITIES.
Find your model sweet spot.
Make the numbers yours
Standard text API estimate. Include reasoning tokens in output.
A scenario estimate, not a measured task bill. Excludes cache discounts, tool fees, retries and provider premiums. Workloads exceeding a catalog context limit receive no cost or Value rating. Advertised context is not a guarantee of retrieval quality.
THE MARKET MAP
Capability meets cost.
Sweet-spot zoneEfficient frontierTap a dot to inspect
INSPECTINGDeepSeek V4.1 Flash ↗
78.9Agent Fit$5.4your workloadYOUR FAST SHORTLIST
How we choose ↗Closest to the sweet spot
- 1DeepSeek V4.1 Flash ↗78.9 Agent Fit · $5.4
- 2Muse Spark 1.3 ↗79.3 Agent Fit · $21
- 3DeepSeek V4 Flash Vision Exp ↗75.1 Agent Fit · $3.45
Shaded zone: 75.2+ points and $21.9 or less. Numbers mark efficient models in or nearest this zone. This shortlist changes with the metric, workload and filters.
More capability.
Less budget.
Practical Value · General agent
70% Agent Fit + 30% affordability. Rankings use your workload; filters do not change the reference cohort.
THE EVIDENCE, SIDE BY SIDE
AI model comparison table
58 of 58 models · select up to 4 to comparePrices in USD · tap ? for instant explanations
01
DeepSeek V4.1 Flash ↗
DeepSeek · max
Agent Fit 78.9
Value 80.6
Your cost $5.4
1.05M contextOpen weightsTool calling
Benchmarks & API pricing +
- LiveBench
- 81.1
- Evidence Consensus
- Not enough evidence
- Input / output per 1M tokens
- $0.3 / $1.2
- Price observation
- OpenRouter catalog
02
DeepSeek V4 Flash Vision Exp ↗
DeepSeek · experimental · source default
Agent Fit 75.1
Value 79.9
Your cost $3.45
1.05M contextOpen weightsTool calling
Benchmarks & API pricing +
- LiveBench
- 76.8
- Evidence Consensus
- Not enough evidence
- Input / output per 1M tokens
- $0.2156 / $0.6468
- Price observation
- OpenRouter catalog
03
DeepSeek V4 Flash 0731 ↗
DeepSeek · source default
Agent Fit 71.1
Value 78.1
Your cost $2.64
1.05M contextOpen weightsTool calling
Benchmarks & API pricing +
- LiveBench
- 74.2
- Evidence Consensus
- Not enough evidence
- Input / output per 1M tokens
- $0.0077 / $1.28
- Price observation
- OpenRouter catalog
04
GPT-6 Luna ↗
OpenAI · max
Agent Fit 67.7
Value 76.8
Your cost $2
1.05M contextTool calling
Benchmarks & API pricing +
- LiveBench
- 72.0
- Evidence Consensus
- Not enough evidence
- Input / output per 1M tokens
- $0.1 / $0.5
- Price observation
- Direct API
05
GPT-5.6 Luna ↗
OpenAI · source ID max; display xhigh
Agent Fit 70.4
Value 76.0
Your cost $4.4
1.05M contextTool calling
Benchmarks & API pricing +
- LiveBench
- 73.6
- Evidence Consensus
- Not enough evidence
- Input / output per 1M tokens
- $0.2 / $1.2
- Price observation
- Direct API
06
GLM-5.3 Flash ↗
Z.ai · source default
Agent Fit 67.2
Value 76.0
Your cost $2.5
1.05M contextOpen weightsTool calling
Benchmarks & API pricing +
- LiveBench
- 71.6
- Evidence Consensus
- Not enough evidence
- Input / output per 1M tokens
- $0.15 / $0.5
- Price observation
- OpenRouter catalog
07
DeepSeek V4 Pro ↗
DeepSeek · source default
Agent Fit 67.1
Value 74.8
Your cost $2.92
1.05M contextOpen weightsTool calling
Benchmarks & API pricing +
- LiveBench
- 71.6
- Evidence Consensus
- Not enough evidence
- Input / output per 1M tokens
- $0.2088 / $0.4176
- Price observation
- OpenRouter catalog
08
GPT-5.4 nano ↗
OpenAI · xhigh
Agent Fit 67.7
Value 73.6
Your cost $4.5
400K contextTool calling
Benchmarks & API pricing +
- LiveBench
- 69.6
- Evidence Consensus
- Not enough evidence
- Input / output per 1M tokens
- $0.2 / $1.25
- Price observation
- Direct API
09
Qwen3.8 27B ↗
Alibaba · source default
Agent Fit 73.5
Value 73.3
Your cost $10.2
1.00M contextOpen weightsTool calling
Benchmarks & API pricing +
- LiveBench
- 75.3
- Evidence Consensus
- Not enough evidence
- Input / output per 1M tokens
- $0.42 / $3
- Price observation
- OpenRouter catalog
10
DeepSeek V4 Flash ↗
DeepSeek · source default
Agent Fit 61.6
Value 73.1
Your cost $0.392
1.05M contextOpen weightsTool calling
Benchmarks & API pricing +
- LiveBench
- 65.5
- Evidence Consensus
- Not enough evidence
- Input / output per 1M tokens
- $0.028 / $0.056
- Price observation
- OpenRouter catalog
11
DeepSeek V4 Pro 0813 ↗
DeepSeek · source default
Agent Fit 73.3
Value 72.6
Your cost $10.56
1.05M contextOpen weightsTool calling
Benchmarks & API pricing +
- LiveBench
- 77.4
- Evidence Consensus
- Not enough evidence
- Input / output per 1M tokens
- $0.66 / $1.98
- Price observation
- OpenRouter catalog
12
Gemini 3.7 Flash ↗
Google · high
Agent Fit 76.1
Value 72.3
Your cost $15
1.05M contextTool calling
Benchmarks & API pricing +
- LiveBench
- 78.8
- Evidence Consensus
- 50.0
- Input / output per 1M tokens
- $0.75 / $3.75
- Price observation
- OpenRouter catalog
| Compare | Input / outputUSD / 1M tokens | ||||||||
|---|---|---|---|---|---|---|---|---|---|
01 DeepSeek V4.1 Flash ↗DeepSeek · maxOpen weights | 78.9 | 80.6 | — | 81.1 | — | $5.4 | 1.05M | $0.3 / $1.2OpenRouter catalog | |
02 DeepSeek V4 Flash Vision Exp ↗DeepSeek · experimental · source defaultOpen weightsPreview | 75.1 | 79.9 | — | 76.8 | — | $3.45 | 1.05M | $0.2156 / $0.6468OpenRouter catalog | |
03 DeepSeek V4 Flash 0731 ↗DeepSeek · source defaultOpen weights | 71.1 | 78.1 | — | 74.2 | — | $2.64 | 1.05M | $0.0077 / $1.28OpenRouter catalog | |
04 GPT-6 Luna ↗OpenAI · max | 67.7 | 76.8 | — | 72.0 | — | $2 | 1.05M | $0.1 / $0.5Direct API | |
05 GPT-5.6 Luna ↗OpenAI · source ID max; display xhighConfiguration discrepancy | 70.4 | 76.0 | — | 73.6 | — | $4.4 | 1.05M | $0.2 / $1.2Direct API | |
06 GLM-5.3 Flash ↗Z.ai · source defaultOpen weights | 67.2 | 76.0 | — | 71.6 | — | $2.5 | 1.05M | $0.15 / $0.5OpenRouter catalog | |
07 DeepSeek V4 Pro ↗DeepSeek · source defaultOpen weights | 67.1 | 74.8 | — | 71.6 | — | $2.92 | 1.05M | $0.2088 / $0.4176OpenRouter catalog | |
08 GPT-5.4 nano ↗OpenAI · xhigh | 67.7 | 73.6 | — | 69.6 | — | $4.5 | 400K | $0.2 / $1.25Direct API | |
09 Qwen3.8 27B ↗Alibaba · source defaultOpen weights | 73.5 | 73.3 | — | 75.3 | — | $10.2 | 1.00M | $0.42 / $3OpenRouter catalog | |
10 DeepSeek V4 Flash ↗DeepSeek · source defaultOpen weights | 61.6 | 73.1 | — | 65.5 | — | $0.392 | 1.05M | $0.028 / $0.056OpenRouter catalog | |
11 DeepSeek V4 Pro 0813 ↗DeepSeek · source defaultOpen weights | 73.3 | 72.6 | — | 77.4 | — | $10.56 | 1.05M | $0.66 / $1.98OpenRouter catalog | |
12 Gemini 3.7 Flash ↗Google · high | 76.1 | 72.3 | 50.02 sources | 78.8 | 1,488±5 · preliminary | $15 | 1.05M | $0.75 / $3.75OpenRouter catalog | |
13 Muse Spark 1.3 ↗Meta · xhigh | 79.3 | 71.3 | — | 81.6 | — | $21 | 1.05M | $1.25 / $4.25OpenRouter catalog | |
14 Kimi K2.6 ↗Moonshot AI · thinkingOpen weights | 66.9 | 70.8 | — | 70.5 | — | $8 | 262K | $0.4342 / $1.83OpenRouter catalog | |
15 Gemini 3.8 Flash ↗Google · high | 73.3 | 70.4 | 55.62 sources | 75.8 | 1,495±5 · preliminary | $15 | 1.05M | $0.75 / $3.75OpenRouter catalog | |
16 MiniMax M3 ↗MiniMax · source default | 63.1 | 69.5 | — | 67.3 | — | $5.4 | 1.05M | $0.3 / $1.2OpenRouter catalog | |
17 Qwen3.6 Plus ↗Alibaba · source default | 63.9 | 69.3 | — | 68.9 | — | $7.15 | 1.00M | $0.325 / $1.95OpenRouter catalog | |
18 Muse Spark 1.2 ↗Meta · xhigh | 76.3 | 69.2 | 55.62 sources | 78.0 | 1,494±9 | $21 | 1.05M | $1.25 / $4.25OpenRouter catalog | |
19 GLM-5.2 ↗Z.ai · source defaultOpen weights | 68.5 | 68.7 | — | 73.2 | — | $12.08 | 1.05M | $0.41 / $3.99OpenRouter catalog | |
20 Gemini 3.6 Flash ↗Google · high | 70.3 | 68.3 | 13.92 sources | 73.6 | 1,483±4 | $15 | 1.05M | $0.75 / $3.75OpenRouter catalog | |
21 Muse Spark 1.1 ↗Meta · source ID xhigh; display highConfiguration discrepancy | 74.0 | 67.6 | — | 75.3 | — | $21 | 1.05M | $1.25 / $4.25OpenRouter catalog | |
22 Nemotron 3 Ultra ↗NVIDIA · 550B A55B · source defaultOpen weights | 63.8 | 67.5 | — | 67.4 | — | $9.4 | 262K | $0.5 / $2.2OpenRouter catalog | |
23 GLM-5.3 ↗Z.ai · source defaultOpen weights | 73.7 | 66.3 | — | 76.1 | — | $22.8 | 1.05M | $1.4 / $4.4OpenRouter catalog | |
24 Grok 4.6 ↗xAI · source default | 75.3 | 65.8 | — | 78.0 | — | $32 | 500K | $2 / $6OpenRouter catalog | |
25 Kimi K2.7 Code ↗Moonshot AI · source defaultOpen weights | 64.8 | 65.6 | — | 68.4 | — | $13.41 | 262K | $0.6712 / $3.35OpenRouter catalog | |
26 Inkling ↗Thinking Machines · xhighOpen weights | 68.9 | 65.2 | — | 71.9 | — | $17.6 | 524K | $0.95 / $4.05OpenRouter catalog | |
27 Gemini 3.5 Flash-Lite ↗Google · high | 59.5 | 65.1 | — | 63.9 | — | $8 | 1.05M | $0.3 / $2.5OpenRouter catalog | |
28 GPT-6.1 Sol ↗OpenAI · max | 77.7 | 65.0 | 52.82 sources | 81.6 | 1,483±11 | $40 | 1.05M | $2 / $10Direct API | |
29 Grok 4.7 ↗xAI · xhigh | 73.7 | 64.7 | — | 77.4 | — | $32 | 500K | $2 / $6OpenRouter catalog | |
30 Qwen3.6 27B ↗Alibaba · source defaultOpen weights | 60.0 | 64.4 | — | 64.0 | — | $9.6 | 262K | $0.32 / $3.2OpenRouter catalog | |
31 Grok 4.5 ↗xAI · source default | 73.1 | 64.3 | — | 75.8 | — | $32 | 500K | $2 / $6OpenRouter catalog | |
32 Qwen3.7 Max ↗Alibaba · source default | 70.4 | 63.5 | — | 73.1 | — | $23.6 | 1.00M | $1.48 / $4.43OpenRouter catalog | |
33 GPT-6 Sol ↗OpenAI · max | 74.7 | 62.9 | — | 79.2 | — | $40 | 1.05M | $2 / $10Direct API | |
34 Claude Sonnet 5 ↗Anthropic · xhigh | 73.3 | 61.9 | — | 76.0 | — | $40 | 1.00M | $2 / $10Direct API | |
35 GPT-5.4 mini ↗OpenAI · xhigh | 62.5 | 61.7 | — | 66.4 | — | $16.5 | 400K | $0.75 / $4.5Direct API | |
36 Gemini 3.5 Flash ↗Google · high | 70.8 | 61.6 | 13.92 sources | 74.6 | 1,477±4 | $33 | 1.05M | $1.5 / $9OpenRouter catalog | |
37 Kimi K3 ↗Moonshot AI · source defaultOpen weights | 77.4 | 61.3 | — | 79.2 | — | $54 | 1.05M | $2.7 / $13.5OpenRouter catalog | |
38 GPT-5.6 Terra ↗OpenAI · source ID max; display xhighConfiguration discrepancy | 74.1 | 60.8 | — | 77.9 | — | $44 | 1.05M | $2 / $12Direct API | |
39 Claude Opus 5.5 ↗Anthropic · max | 79.4 | 60.8 | — | 83.2 | — | $80 | 1.00M | $4 / $20Direct API | |
40 Claude Sonnet 5.5 ↗Anthropic · max | 71.0 | 60.3 | — | 75.7 | — | $40 | 1.00M | $2 / $10Direct API | |
41 Gemini 3.1 Pro Preview ↗Google · highPreview | 73.2 | 60.3 | — | 77.0 | — | $44 | 1.05M | $2 / $12OpenRouter catalog | |
42 GPT-5.6 Sol ↗OpenAI · source ID max; display xhighConfiguration discrepancy | 77.1 | 59.1 | — | 81.1 | — | $80 | 1.05M | $4 / $20Direct API | |
43 GPT-5.4 ↗OpenAI · xhigh | 74.4 | 58.6 | — | 78.0 | — | $55 | 1.05M | $2.5 / $15Direct API | |
44 GPT-5.2 Codex ↗OpenAI · source default | 69.9 | 56.8 | — | 74.0 | — | $45.5 | 400K | $1.75 / $14OpenRouter catalog | |
45 GPT-5.2 ↗OpenAI · high · 2025-12-11 | 69.8 | 56.8 | — | 74.6 | — | $45.5 | 400K | $1.75 / $14OpenRouter catalog | |
46 Grok 4.3 ↗xAI · source default | 56.0 | 56.7 | — | 62.2 | — | $17.5 | 1.00M | $1.25 / $2.5OpenRouter catalog | |
47 Claude Fable 5.1 ↗Anthropic · max | 79.6 | 56.3 | 94.42 sources | 83.4 | 1,501±6 | $200 | 1.00M | $10 / $50Direct API | |
48 Claude Opus 5 ↗Anthropic · max | 75.7 | 56.2 | 61.12 sources | 80.1 | 1,489±5 | $100 | 1.00M | $5 / $25Direct API | |
49 Claude Fable 5 ↗Anthropic · source ID max; display xhighConfiguration discrepancy | 79.0 | 55.8 | — | 83.0 | — | $200 | 1.00M | $10 / $50Direct API | |
50 GPT-6 Astra ↗OpenAI · max | 78.6 | 55.6 | 47.22 sources | 82.2 | 1,477±7 | $200 | 1.05M | $10 / $50Direct API | |
51 GPT-5.5 ↗OpenAI · xhigh | 75.8 | 54.7 | — | 80.2 | — | $110 | 1.05M | $5 / $30Direct API | |
52 Claude Sonnet 4.6 ↗Anthropic · medium · adaptive thinking | 69.4 | 54.6 | — | 73.0 | — | $60 | 1.00M | $3 / $15Direct API | |
53 Claude Opus 4.8 ↗Anthropic · max | 73.0 | 54.3 | — | 76.2 | — | $100 | 1.00M | $5 / $25Direct API | |
54 Claude Opus 4.7 ↗Anthropic · xhigh | 72.9 | 54.3 | — | 76.5 | — | $100 | 1.00M | $5 / $25Direct API | |
55 Claude Opus 4.6 ↗Anthropic · high · adaptive thinking | 70.5 | 52.6 | 55.62 sources | 74.5 | 1,505±3 | $100 | 1.00M | $5 / $25Direct API | |
56 Claude Opus 4.5 ↗Anthropic · high · 64K thinking | 66.7 | 50.0 | — | 72.6 | — | $100 | 200K | $5 / $25Direct API | |
57 Qwen3.8 Flash Next ↗Alibaba · source defaultOpen weights | 76.2 | — | — | 76.2 | — | — | Not verified | Not verified | |
58 Qwen3.8 Max ↗Alibaba · source default · version mapping unconfirmed | 77.0 | — | — | 78.5 | — | — | Not verified | Not verified |
— means missing comparable evidence, never zero. Consensus covers only 10 matched configurations. Model IDs, reasoning settings, source links and pricing differences are available on every model page. Open weights does not imply unrestricted commercial use.
LOOK BEYOND ONE NUMBER
Your shortlist, under the microscope.
Claude Opus 5.5GPT-6 SolDeepSeek V4.1 Flash
| Measure | Claude Opus 5.5max | GPT-6 Solmax | DeepSeek V4.1 Flashmax |
|---|---|---|---|
| Agent Fit | 79.4 | 74.7 | 78.9 |
| Practical Value | 60.8 | 62.9 | 80.6 |
| Evidence Consensus | Insufficient evidence | Insufficient evidence | Insufficient evidence |
| Workload cost | $80 | $40 | $5.4 |
| Context window | 1.00M | 1.05M | 1.05M |
| Tool calling | Supported | Supported | Supported |
| Reasoning | 92.2 | 88.7 | 86.7 |
| Coding | 89.3 | 81.8 | 80.0 |
| Agentic coding | 71.7 | 52.9 | 77.3 |
| Mathematics | 97.1 | 96.4 | 93.3 |
| Data analysis | 80.3 | 81.2 | 79.3 |
| Language | 86.3 | 85.3 | 81.2 |
| Instruction following | 65.7 | 68.6 | 70.0 |
| Input / output per 1M | $4 / $20 | $2 / $10 | $0.3 / $1.2 |