1<!DOCTYPE html><!--RW1gn8FFwfPCFEPOXxt4k--><html lang="en" class="dark"><head><meta charSet="utf-8"/><meta name="viewport" content="width=device-width, initial-scale=1"/><link rel="preload" href="/_next/static/media/797e433ab948586e-s.p.29207c2f.woff2" as="font" crossorigin="" type="font/woff2"/><link rel="preload" href="/_next/static/media/caa3a2e1cccd8315-s.p.3b6cae6d.woff2" as="font" crossorigin="" type="font/woff2"/><link rel="stylesheet" href="/_next/static/chunks/29c48147fb5b609b.css?dpl=dpl_AWxxCfXD9UY7YUK42akeB8WWH9fr" data-precedence="next"/><link rel="preload" as="script" fetchPriority="low" href="/_next/static/chunks/8e44e41d890b5e5b.js?dpl=dpl_AWxxCfXD9UY7YUK42akeB8WWH9fr"/>
1<script src="/_next/static/chunks/806bdb8e4a6a9b95.js?dpl=dpl_AWxxCfXD9UY7YUK42akeB8WWH9fr" async=""></script>
1<script src="/_next/static/chunks/7296d4f93dc2eaa0.js?dpl=dpl_AWxxCfXD9UY7YUK42akeB8WWH9fr" async=""></script>
1<script src="/_next/static/chunks/69be39811437728d.js?dpl=dpl_AWxxCfXD9UY7YUK42akeB8WWH9fr" async=""></script>
1<script src="/_next/static/chunks/turbopack-e220d3a1ee97c35e.js?dpl=dpl_AWxxCfXD9UY7YUK42akeB8WWH9fr" async=""></script>
1<script src="/_next/static/chunks/ff1a16fafef87110.js?dpl=dpl_AWxxCfXD9UY7YUK42akeB8WWH9fr" async=""></script>
1<script src="/_next/static/chunks/d2be314c3ece3fbe.js?dpl=dpl_AWxxCfXD9UY7YUK42akeB8WWH9fr" async=""></script>
1<script src="/_next/static/chunks/fcb51f43b14e3807.js?dpl=dpl_AWxxCfXD9UY7YUK42akeB8WWH9fr" async=""></script>
1<meta name="next-size-adjust" content=""/><title>Luminal - Inference at the Speed of Light</title><meta name="description" content="Luminal is an AI inference compiler that compiles and optimizes AI models for GPUs and ASICs, delivering the fastest, highest throughput inference in the world."/><meta name="keywords" content="Luminal,Luminal AI,AI inference,AI compiler,LLM inference,machine learning inference,model optimization,GPU inference,ASIC inference,AI inference engine,high throughput inference,low latency inference,AI model deployment,inference optimization,AI infrastructure"/><meta name="robots" content="index, follow"/><link rel="canonical" href="https://luminal.com"/><meta property="og:title" content="Luminal - Inference at the Speed of Light"/><meta property="og:description" content="Luminal is an AI inference compiler that compiles and optimizes AI models for GPUs and ASICs, delivering the fastest, highest throughput inference in the world."/><meta property="og:url" content="https://luminal.com"/><meta property="og:site_name" content="Luminal"/><meta property="og:type" content="website"/><meta name="twitter:card" content="summary_large_image"/><meta name="twitter:title" content="Luminal - Inference at the Speed of Light"/><meta name="twitter:description" content="Luminal is an AI inference compiler that compiles and optimizes AI models for GPUs and ASICs, delivering the fastest, highest throughput inference in the world."/><link rel="icon" href="/icon.svg?icon.92a21eff.svg" sizes="any" type="image/svg+xml"/>
1<script src="/_next/static/chunks/a6dad97d9634a72d.js?dpl=dpl_AWxxCfXD9UY7YUK42akeB8WWH9fr" noModule=""></script>
1</head><body class="geist_a71539c9-module__T19VSG__variable geist_mono_8d43a2aa-module__8Li5zG__variable antialiased bg-[#030712] text-gray-100"><div hidden=""><!--$--><!--/$--></div><div class="min-h-screen pb-16"><nav class="fixed top-0 left-0 right-0 z-50 bg-[#080F11]/90 backdrop-blur-sm border-b border-[#2D3335]"><div class="max-w-[1200px] mx-auto px-6 h-14 flex items-center justify-between"><a class="flex items-center gap-2.5" href="/"><span class="flex items-center gap-2.5 font-mono text-[#2faa6e] text-sm font-medium tracking-wide"><span class="relative w-4 h-4 rounded-full bg-[#1a5c3a] flex items-center justify-center"><span class="w-2 h-2 rounded-full bg-[#2faa6e]"></span></span>LUMINAL</span></a><div class="hidden sm:flex items-center gap-6"><a class="label hover:text-[#D7D8D9] transition-colors" href="/report">Benchmarks</a><a class="label hover:text-[#D7D8D9] transition-colors" href="/calculator">Calculator</a><a href="https://blog.luminal.com" target="_blank" rel="noopener noreferrer" class="label hover:text-[#D7D8D9] transition-colors">Blog</a><a class="label hover:text-[#D7D8D9] transition-colors" href="/careers">Careers</a><button class="label hover:text-[#D7D8D9] transition-colors cursor-pointer">Contact</button><a href="https://github.com/luminal-ai/luminal" target="_blank" rel="noopener noreferrer" class="text-[#7e8385] hover:text-[#D7D8D9] transition-colors"><svg class="w-4 h-4" viewBox="0 0 24 24" fill="currentColor"><path d="M12 0c-6.626 0-12 5.373-12 12 0 5.302 3.438 9.8 8.207 11.387.599.111.793-.261.793-.577v-2.234c-3.338.726-4.033-1.416-4.033-1.416-.546-1.387-1.333-1.756-1.333-1.756-1.089-.745.083-.729.083-.729 1.205.084 1.839 1.237 1.839 1.237 1.07 1.834 2.807 1.304 3.492.997.107-.775.418-1.305.762-1.604-2.665-.305-5.467-1.334-5.467-5.931 0-1.311.469-2.381 1.236-3.221-.124-.303-.535-1.524.117-3.176 0 0 1.008-.322 3.301 1.23.957-.266 1.983-.399 3.003-.404 1.02.005 2.047.138 3.006.404 2.291-1.552 3.297-1.23 3.297-1.23.653 1.653.242 2.874.118 3.176.77.84 1.235 1.911 1.235 3.221 0 4.609-2.807 5.624-5.479 5.921.43.372.823 1.102.823 2.222v3.293c0 .319.192.694.801.576 4.765-1.589 8.199-6.086 8.199-11.386 0-6.627-5.373-12-12-12z"></path></svg></a><button class="font-mono text-[11px] tracking-[0.1em] uppercase text-[#2faa6e] border border-[#2faa6e]/30 px-4 py-1.5 hover:bg-[#2faa6e] hover:text-[#080F11] transition-colors rounded-sm cursor-pointer">Get Started</button></div><button class="sm:hidden text-[#A1A4A5] hover:text-[#D7D8D9] transition-colors cursor-pointer" aria-label="Toggle menu"><svg class="w-5 h-5" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round"><line x1="4" y1="7" x2="20" y2="7"></line><line x1="4" y1="12" x2="20" y2="12"></line><line x1="4" y1="17" x2="20" y2="17"></line></svg></button></div></nav><section class="pt-14"><div class="max-w-[1200px] mx-auto px-0 sm:px-6"><div class="relative sm:border-x border-[#2D3335] border-x-0 min-h-[calc(100vh-56px)] flex flex-col justify-center py-24 overflow-hidden"><div class="absolute inset-0 pointer-events-none min-w-[1200px]" aria-hidden="true"><div class="flow-line" style="top:22%;--delay:-1.2s;--duration:4.8s;--opacity:0.44;--pattern:linear-gradient(90deg, var(--accent-800) 0px, var(--accent-800) 50px, transparent 50px, transparent 56px, var(--accent-800) 56px, var(--accent-800) 95px, transparent 95px, transparent 102px, var(--accent-800) 102px, var(--accent-800) 130px, transparent 130px, transparent 135px);transform:rotate(-1.5deg) scaleX(1.3)"></div><div class="flow-line hidden md:block" style="top:26%;--delay:-2.8s;--duration:3.5s;--opacity:0.34;--pattern:linear-gradient(90deg, var(--accent-800) 0px, var(--accent-800) 38px, transparent 38px, transparent 44px, var(--accent-800) 44px, var(--accent-800) 110px, transparent 110px, transparent 115px, var(--accent-800) 115px, var(--accent-800) 148px, transparent 148px, transparent 155px);transform:rotate(-1.5deg) scaleX(1.3)"></div><div class="flow-line" style="top:30%;--delay:-0.5s;--duration:4.2s;--opacity:0.50;--pattern:linear-gradient(90deg, var(--accent-800) 0px, var(--accent-800) 70px, transparent 70px, transparent 77px, var(--accent-800) 77px, var(--accent-800) 105px, transparent 105px, transparent 110px, var(--accent-800) 110px, var(--accent-800) 160px, transparent 160px, transparent 168px);transform:rotate(-1.5deg) scaleX(1.3)"></div><div class="flow-line hidden md:block" style="top:34%;--delay:-1.9s;--duration:3s;--opacity:0.44;--pattern:linear-gradient(90deg, var(--accent-800) 0px, var(--accent-800) 45px, transparent 45px, transparent 50px, var(--accent-800) 50px, var(--accent-800) 88px, transparent 88px, transparent 95px, var(--accent-800) 95px, var(--accent-800) 140px, transparent 140px, transparent 146px);transform:rotate(-1.5deg) scaleX(1.3)"></div><div class="flow-line" style="top:38%;--delay:-3.4s;--duration:3.8s;--opacity:0.38;--pattern:linear-gradient(90deg, var(--accent-800) 0px, var(--accent-800) 62px, transparent 62px, transparent 68px, var(--accent-800) 68px, var(--accent-800) 120px, transparent 120px, transparent 126px, var(--accent-800) 126px, var(--accent-800) 155px, transparent 155px, transparent 162px);transform:rotate(-1.5deg) scaleX(1.3)"></div><div class="flow-line hidden md:block" style="top:42%;--delay:-0.8s;--duration:3.2s;--opacity:0.50;--pattern:linear-gradient(90deg, var(--accent-800) 0px, var(--accent-800) 55px, transparent 55px, transparent 60px, var(--accent-800) 60px, var(--accent-800) 98px, transparent 98px, transparent 105px, var(--accent-800) 105px, var(--accent-800) 165px, transparent 165px, transparent 170px);transform:rotate(-1.5deg) scaleX(1.3)"></div><div class="flow-line" style="top:46%;--delay:-2.3s;--duration:4.5s;--opacity:0.48;--pattern:linear-gradient(90deg, var(--accent-800) 0px, var(--accent-800) 80px, transparent 80px, transparent 86px, var(--accent-800) 86px, var(--accent-800) 115px, transparent 115px, transparent 122px, var(--accent-800) 122px, var(--accent-800) 172px, transparent 172px, transparent 178px);transform:rotate(-1.5deg) scaleX(1.3)"></div><div class="flow-line hidden md:block" style="top:50%;
1--delay:-1.6s;--duration:2.8s;--opacity:0.55;--pattern:linear-gradient(90deg, var(--accent-800) 0px, var(--accent-800) 42px, transparent 42px, transparent 48px, var(--accent-800) 48px, var(--accent-800) 105px, transparent 105px, transparent 110px, var(--accent-800) 110px, var(--accent-800) 138px, transparent 138px, transparent 145px);transform:rotate(-1.5deg) scaleX(1.3)"></div><div class="flow-line" style="top:54%;--delay:-3.1s;--duration:3.6s;--opacity:0.40;--pattern:linear-gradient(90deg, var(--accent-800) 0px, var(--accent-800) 68px, transparent 68px, transparent 74px, var(--accent-800) 74px, var(--accent-800) 130px, transparent 130px, transparent 136px, var(--accent-800) 136px, var(--accent-800) 158px, transparent 158px, transparent 165px);transform:rotate(-1.5deg) scaleX(1.3)"></div><div class="flow-line hidden md:block" style="top:58%;--delay:-0.3s;--duration:4s;--opacity:0.52;--pattern:linear-gradient(90deg, var(--accent-800) 0px, var(--accent-800) 35px, transparent 35px, transparent 42px, var(--accent-800) 42px, var(--accent-800) 92px, transparent 92px, transparent 97px, var(--accent-800) 97px, var(--accent-800) 145px, transparent 145px, transparent 152px);transform:rotate(-1.5deg) scaleX(1.3)"></div><div class="flow-line" style="top:62%;--delay:-2.5s;--duration:3.3s;--opacity:0.48;--pattern:linear-gradient(90deg, var(--accent-800) 0px, var(--accent-800) 75px, transparent 75px, transparent 80px, var(--accent-800) 80px, var(--accent-800) 112px, transparent 112px, transparent 118px, var(--accent-800) 118px, var(--accent-800) 170px, transparent 170px, transparent 176px);transform:rotate(-1.5deg) scaleX(1.3)"></div><div class="flow-line hidden md:block" style="top:66%;--delay:-1.7s;--duration:4.4s;--opacity:0.26;--pattern:linear-gradient(90deg, var(--accent-800) 0px, var(--accent-800) 58px, transparent 58px, transparent 65px, var(--accent-800) 65px, var(--accent-800) 100px, transparent 100px, transparent 105px, var(--accent-800) 105px, var(--accent-800) 150px, transparent 150px, transparent 158px);transform:rotate(-1.5deg) scaleX(1.3)"></div><div class="flow-line" style="top:70%;--delay:-0.9s;--duration:3.1s;--opacity:0.52;--pattern:linear-gradient(90deg, var(--accent-800) 0px, var(--accent-800) 48px, transparent 48px, transparent 54px, var(--accent-800) 54px, var(--accent-800) 118px, transparent 118px, transparent 123px, var(--accent-800) 123px, var(--accent-800) 155px, transparent 155px, transparent 162px);transform:rotate(-1.5deg) scaleX(1.3)"></div><div class="flow-line hidden md:block" style="top:74%;--delay:-2.1s;--duration:3.7s;--opacity:0.36;--pattern:linear-gradient(90deg, var(--accent-800) 0px, var(--accent-800) 65px, transparent 65px, transparent 72px, var(--accent-800) 72px, var(--accent-800) 108px, transparent 108px, transparent 114px, var(--accent-800) 114px, var(--accent-800) 162px, transparent 162px, transparent 168px);transform:rotate(-1.5deg) scaleX(1.3)"></div><div class="flow-line" style="top:78%;--delay:-3.6s;--duration:4.1s;--opacity:0.42;--pattern:linear-gradient(90deg, var(--accent-800) 0px, var(--accent-800) 40px, transparent 40px, transparent 46px, var(--accent-800) 46px, var(--accent-800) 95px, transparent 95px, transparent 100px, var(--accent-800) 100px, var(--accent-800) 142px, transparent 142px, transparent 150px);transform:rotate(-1.5deg) scaleX(1.3)"></div></div><div class="relative z-10 px-8 md:px-16 max-w-3xl"><h1 class="text-4xl sm:text-5xl md:text-6xl font-medium text-[#D7D8D9] tracking-tight leading-[1.1] mb-6">Inference at the<br/><span class="text-[#2faa6e]">speed of light</span><span class="text-[#2faa6e]">.</span></h1><p class="text-lg text-[#A1A4A5] max-w-xl mb-12 leading-relaxed">Luminal compiles AI models to give you the fastest, highest throughput inference in the world.</p><div class="flex flex-col sm:flex-row items-start sm:items-center gap-4"><button class="inline-flex items-center gap-2 bg-[#2faa6e] text-[#080F11] border border-[#2faa6e] px-6 py-2.5 text-sm font-medium font-mono uppercase tracking-wider hover:bg-transparent hover:text-[#2faa6e] transition-all duration-200 rounded-sm cursor-pointer">
1Get Started<svg class="w-3.5 h-3.5" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"><line x1="5" y1="12" x2="19" y2="12"></line><polyline points="12 5 19 12 12 19"></polyline></svg></button><a href="https://github.com/luminal-ai/luminal" target="_blank" rel="noopener noreferrer" class="inline-flex items-center gap-2 border border-[#2D3335] text-[#A1A4A5] px-6 py-2.5 text-sm font-medium font-mono uppercase tracking-wider hover:border-[#404647] hover:text-[#D7D8D9] transition-colors rounded-sm"><svg class="w-4 h-4" viewBox="0 0 24 24" fill="currentColor"><path d="M12 0c-6.626 0-12 5.373-12 12 0 5.302 3.438 9.8 8.207 11.387.599.111.793-.261.793-.577v-2.234c-3.338.726-4.033-1.416-4.033-1.416-.546-1.387-1.333-1.756-1.333-1.756-1.089-.745.083-.729.083-.729 1.205.084 1.839 1.237 1.839 1.237 1.07 1.834 2.807 1.304 3.492.997.107-.775.418-1.305.762-1.604-2.665-.305-5.467-1.334-5.467-5.931 0-1.311.469-2.381 1.236-3.221-.124-.303-.535-1.524.117-3.176 0 0 1.008-.322 3.301 1.23.957-.266 1.983-.399 3.003-.404 1.02.005 2.047.138 3.006.404 2.291-1.552 3.297-1.23 3.297-1.23.653 1.653.242 2.874.118 3.176.77.84 1.235 1.911 1.235 3.221 0 4.609-2.807 5.624-5.479 5.921.43.372.823 1.102.823 2.222v3.293c0 .319.192.694.801.576 4.765-1.589 8.199-6.086 8.199-11.386 0-6.627-5.373-12-12-12z"></path></svg>GitHub</a></div></div></div></div></section><section><div class="max-w-[1200px] mx-auto px-0 sm:px-6"><div class="sm:border-x border-[#2D3335] border-x-0"><div class="border-b border-[#2D3335] px-8 md:px-16 py-4"><span class="label-accent">The Compiler</span></div><div class="px-8 md:px-16 py-16"><h2 class="text-2xl sm:text-3xl font-medium text-[#D7D8D9] tracking-tight mb-4">Compiled inference, not interpreted</h2><p class="text-[#A1A4A5] leading-relaxed mb-16 max-w-2xl">Unlike runtime inference engines that interpret models dynamically, Luminal compiles your model ahead of time into optimized native code for GPUs and ASICs, eliminating every layer of overhead.</p><div class="border border-[#2D3335] mb-16"><div class="border-b border-[#2D3335] px-5 py-3"><span class="label">Compilation Pipeline</span></div><div class="p-6 md:p-8"><div class="hidden md:flex items-center w-full"><div class="flex items-center flex-1 relative"><div class="absolute top-1/2 left-[50px] right-[50px] h-px bg-[#1C2225] -translate-y-1/2 overflow-hidden"><div class="pipeline-pulse" style="animation-delay:0s"></div><div class="pipeline-pulse" style="animation-delay:-1.5s"></div></div><div class="flex items-center"><div class="relative z-10 border border-[#2D3335] bg-[#0d1416] px-4 py-3 text-center min-w-[100px] shrink-0"><div class="font-mono text-xs tracking-wider text-[#2faa6e] mb-1">MODEL</div><div class="text-[10px] text-[#5B5F61] font-mono uppercase tracking-wider">PyTorch / HF</div></div></div><div class="flex items-center flex-1"><div class="flex-1 flex items-center min-w-[24px] text-[#2faa6e]"><div class="flex-1 h-px bg-[#2D3335]"></div><svg class="w-3 h-3 shrink-0 -ml-px" viewBox="0 0 12 12" fill="none" stroke="currentColor" stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round"><polyline points="4 2 8 6 4 10"></polyline></svg></div><div class="relative z-10 border border-[#2D3335] bg-[#0d1416] px-4 py-3 text-center min-w-[100px] shrink-0"><div class="font-mono text-xs tracking-wider text-[#2faa6e] mb-1">GRAPH IR</div><div class="text-[10px] text-[#5B5F61] font-mono uppercase tracking-wider">Dataflow graph</div></div></div><div class="flex items-center flex-1"><div class="flex-1 flex items-center min-w-[24px] text-[#2faa6e]"><div class="flex-1 h-px bg-[#2D3335]"></div><svg class="w-3 h-3 shrink-0 -ml-px" viewBox="0 0 12 12" fill="none" stroke="currentColor" stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round"><polyline points="4 2 8 6 4 10"></polyline></svg></div><div class="relative z-10 border border-[#2D3335] bg-[#0d1416] px-4 py-3 text-center min-w-[100px] shrink-0"><div class="font-mono text-xs tracking-wider text-[#2faa6e] mb-1">OPTIMIZE</div><div class="text-[10px] text-[#5B5F61] font-mono uppercase tracking-wider">Fuse & schedule</div></div></div><div class="flex items-center flex-1 hidden xl:flex"><div class="flex-1 flex items-center min-w-[24px] text-[#2faa6e]"><div class="flex-1 h-px bg-[#2D3335]"></div><svg class="w-3 h-3 shrink-0 -ml-px" viewBox="0 0 12 12" fill="none" stroke="currentColor" stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round"><polyline points="4 2 8 6 4 10"></polyline></svg></div><div class="relative z-10 border border-[#2D3335] bg-[#0d1416] px-4 py-3 text-center min-w-[100px] shrink-0"><div class="font-mono text-xs tracking-wider text-[#2faa6e] mb-1">CODEGEN</div><div class="text-[10px] text-[#5B5F61] font-mono uppercase tracking-wider">Emit kernels</div></div></div><div class="flex items-center flex-1"><div class="flex-1 flex items-center min-w-[24px] text-[#2faa6e]"><div class="flex-1 h-px bg-[#2D3335]"></div><svg class="w-3 h-3 shrink-0 -ml-px" viewBox="0 0 12 12" fill="none" stroke="currentColor" stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round"><polyline points="4 2 8 6 4 10"></polyline></svg></div><div class="relative z-10 border border-[#2D3335] bg-[#0d1416] px-4 py-3 text-center min-w-[100px] shrink-0"><div class="font-mono text-xs tracking-wider text-[#2faa6e] mb-1">GPU / ASIC</div><div class="text-[10px] text-[#5B5F61] font-mono uppercase tracking-wider">Execute</div></div></div></div></div><div class="flex md:hidden flex-col items-center relative"><div class="absolute left-1/2 top-[24px] bottom-[24px] w-px bg-[#1C2225] -translate-x-1/2 overflow-hidden"><div class="pipeline-pulse-v" style="animation-delay:0s"></div><div class="pipeline-pulse-v" style="animation-delay:-1.5s"></div></div><div class="flex flex-col items-center"><div class="relative z-10 border border-[#2D3335] bg-[#0d1416] px-6 py-3 text-center min-w-[140px]"><div class="font-mono text-xs tracking-wider text-[#2faa6e] mb-1">MODEL</div><div class="text-[10px] text-[#5B5F61] font-mono uppercase tracking-wider">PyTorch / HF</div></div></div><div class="flex flex-col items-center"><div class="flex flex-col items-center h-8 text-[#2faa6e]"><div class="flex-1 w-px bg-[#2D3335]"></div><svg class="w-3 h-3 shrink-0 -mt-px" viewBox="0 0 12 12" fill="none" stroke="currentColor" stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round"><polyline points="2 4 6 8 10 4"></polyline></svg></div><div class="relative z-10 border border-[#2D3335] bg-[#0d1416] px-6 py-3 text-center min-w-[140px]"><div class="font-mono text-xs tracking-wider text-[#2faa6e] mb-1">GRAPH IR</div><div class="text-[10px] text-[#5B5F61] font-mono uppercase tracking-wider">Dataflow graph</div></div></div><div class="flex flex-col items-center"><div class="flex flex-col items-center h-8 text-[#2faa6e]"><div class="flex-1 w-px bg-[#2D3335]"></div><svg class="w-3 h-3 shrink-0 -mt-px" viewBox="0 0 12 12" fill="none" stroke="currentColor" stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round"><polyline points="2 4 6 8 10 4"></polyline></svg></div><div class="relative z-10 border border-[#2D3335] bg-[#0d1416] px-6 py-3 text-center min-w-[140px]"><div class="font-mono text-xs tracking-wider text-[#2faa6e] mb-1">OPTIMIZE</div><div class="text-[10px] text-[#5B5F61] font-mono uppercase tracking-wider">Fuse & schedule</div></div></div><div class="flex flex-col items-center"><div class="flex flex-col items-center h-8 text-[#2faa6e]"><div class="flex-1 w-px bg-[#2D3335]"></div><svg class="w-3 h-3 shrink-0 -mt-px" viewBox="0 0 12 12" fill="none" stroke="currentColor" stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round"><polyline points="2 4 6 8 10 4"></polyline></svg></div><div class="relative z-10 border border-[#2D3335] bg-[#0d1416] px-6 py-3 text-center min-w-[140px]"><div class="font-mono text-xs tracking-wider text-[#2faa6e] mb-1">GPU / ASIC</div><div class="text-[10px] text-[#5B5F61] font-mono uppercase tracking-wider">Execute</div></div></div></div></div></div><div class="grid md:grid-cols-3 gap-0"><div class="p-6 border border-[#2D3335] "><div class="flex items-center justify-between mb-4"><h3 class="text-base font-medium text-[#D7D8D9]">Graph-Level IR</h3><span class="font-mono text-xs text-[#404647]">01</span></div><p class="text-sm text-[#7e8385] leading-relaxed">Models are lowered to a minimal graph intermediate representation, a pure dataflow graph with no framework overhead.</p></div><div class="p-6 border border-[#2D3335] md:border-l-0"><div class="flex items-center justify-between mb-4"><h3 class="text-base font-medium text-[#D7D8D9]">Hardware-Aware Optimization</h3><span class="font-mono text-xs text-[#404647]">02</span></div><p class="text-sm text-[#7e8385] leading-relaxed">The compiler applies fusion, tiling, memory planning, and scheduling passes tuned for each target, GPUs and ASICs.</p></div><div class="p-6 border border-[#2D3335] md:border-l-0"><div class="flex items-center justify-between mb-4"><h3 class="text-base font-medium text-[#D7D8D9]">Zero-Overhead Codegen</h3><span class="font-mono text-xs text-[#404647]">03</span></div><p class="text-sm text-[#7e8385] leading-relaxed">Final code is emitted directly to GPU kernels or ASIC instructions with
1no excess runtime overhead.</p></div></div></div></div></div></section><section><div class="max-w-[1200px] mx-auto px-0 sm:px-6"><div class="sm:border-x border-[#2D3335] border-x-0"><div class="border-b border-[#2D3335] px-8 md:px-16 py-4"><span class="label-accent">The Inference OS</span></div><div class="px-8 md:px-16 py-16"><h2 class="text-2xl sm:text-3xl font-medium text-[#D7D8D9] tracking-tight mb-4">Hyperscale Inference OS</h2><p class="text-[#A1A4A5] leading-relaxed mb-16 max-w-2xl">Luminal dynamically schedules and load-balances inference workloads at any scale, from single accelerators up to large clusters of heterogeneous compute nodes, minimizing latency and maximizing throughput by optimizing inference topologies on-the-fly.</p><div class="border border-[#2D3335] mb-16"><div class="border-b border-[#2D3335] px-5 py-3 flex items-center justify-between"><span class="label">Compute Cluster</span><div class="flex items-center gap-4"><div class="flex items-center gap-1.5"><span class="w-2 h-2 rounded-full" style="background-color:#2faa6e"></span><span class="font-mono text-[10px] tracking-wider text-[#5B5F61] uppercase">gpu</span></div><div class="flex items-center gap-1.5"><span class="w-2 h-2 rounded-full" style="background-color:#3b82f6"></span><span class="font-mono text-[10px] tracking-wider text-[#5B5F61] uppercase">asic</span></div><div class="flex items-center gap-1.5"><span class="w-2 h-2 rounded-full" style="background-color:#a855f7"></span><span class="font-mono text-[10px] tracking-wider text-[#5B5F61] uppercase">cpu</span></div></div></div><div class="hidden md:block relative min-h-[480px] p-6"><svg class="absolute inset-0 w-full h-full pointer-events-none z-0" aria-hidden="true"><g>
1<line x1="18%" y1="20%" x2="50%" y2="13%" stroke="#1C2225" stroke-width="1"></line><line x1="18%" y1="20%" x2="50%" y2="13%" stroke="#2faa6e" stroke-width="1" stroke-opacity="0.4" stroke-dasharray="40 80"><animate attributeName="stroke-dashoffset" from="120" to="0" dur="0.4s" repeatCount="indefinite" begin="0s"></animate></line><line x1="18%" y1="20%" x2="50%" y2="13%" stroke="#2faa6e" stroke-width="1" stroke-opacity="0.25" stroke-dasharray="30 90"><animate attributeName="stroke-dashoffset" from="0" to="120" dur="0.4s" repeatCount="indefinite" begin="-0.4s"></animate></line></g><g><line x1="18%" y1="20%" x2="12%" y2="68%" stroke="#1C2225" stroke-width="1"></line><line x1="18%" y1="20%" x2="12%" y2="68%" stroke="#2faa6e" stroke-width="1" stroke-opacity="0.4" stroke-dasharray="40 80"><animate attributeName="stroke-dashoffset" from="120" to="0" dur="0.35s" repeatCount="indefinite" begin="-0.15s"></animate></line><line x1="18%" y1="20%" x2="12%" y2="68%" stroke="#2faa6e" stroke-width="1" stroke-opacity="0.25" stroke-dasharray="30 90"><animate attributeName="stroke-dashoffset" from="0" to="120" dur="0.35s" repeatCount="indefinite" begin="-0.55s"></animate></line></g><g><line x1="50%" y1="13%" x2="84%" y2="22%" stroke="#1C2225" stroke-width="1"></line><line x1="50%" y1="13%" x2="84%" y2="22%" stroke="#2faa6e" stroke-width="1" stroke-opacity="0.4" stroke-dasharray="40 80"><animate attributeName="stroke-dashoffset" from="120" to="0" dur="0.45s" repeatCount="indefinite" begin="-0.25s"></animate></line><line x1="50%" y1="13%" x2="84%" y2="22%" stroke="#2faa6e" stroke-width="1" stroke-opacity="0.25" stroke-dasharray="30 90"><animate attributeName="stroke-dashoffset" from="0" to="120" dur="0.45s" repeatCount="indefinite" begin="-0.65s"></animate></line></g><g><line x1="50%" y1="13%" x2="46%" y2="50%" stroke="#1C2225" stroke-width="1"></line><line x1="50%" y1="13%" x2="46%" y2="50%" stroke="#2faa6e" stroke-width="1" stroke-opacity="0.4" stroke-dasharray="40 80"><animate attributeName="stroke-dashoffset" from="120" to="0" dur="0.35s" repeatCount="indefinite" begin="-0.1s"></animate></line><line x1="50%" y1="13%" x2="46%" y2="50%" stroke="#2faa6e" stroke-width="1" stroke-opacity="0.25" stroke-dasharray="30 90"><animate attributeName="stroke-dashoffset" from="0" to="120" dur="0.35s" repeatCount="indefinite" begin="-0.5s"></animate></line></g><g><line x1="84%" y1="22%" x2="82%" y2="66%" stroke="#1C2225" stroke-width="1"></line><line x1="84%" y1="22%" x2="82%" y2="66%" stroke="#2faa6e" stroke-width="1" stroke-opacity="0.4" stroke-dasharray="40 80"><animate attributeName="stroke-dashoffset" from="120" to="0" dur="0.4s" repeatCount="indefinite" begin="-0.4s"></animate></line><line x1="84%" y1="22%" x2="82%" y2="66%" stroke="#2faa6e" stroke-width="1" stroke-opacity="0.25" stroke-dasharray="30 90"><animate attributeName="stroke-dashoffset" from="0" to="120" dur="0.4s" repeatCount="indefinite" begin="-0.8s"></animate></line></g><g><line x1="12%" y1="68%" x2="46%" y2="50%" stroke="#1C2225" stroke-width="1"></line><line x1="12%" y1="68%" x2="46%" y2="50%" stroke="#2faa6e" stroke-width="1" stroke-opacity="0.4" stroke-dasharray="40 80"><animate attributeName="stroke-dashoffset" from="120" to="0" dur="0.45s" repeatCount="indefinite" begin="-0.2s"></animate></line><line x1="12%" y1="68%" x2="46%" y2="50%" stroke="#2faa6e" stroke-width="1" stroke-opacity="0.25" stroke-dasharray="30 90"><animate attributeName="stroke-dashoffset" from="0" to="120" dur="0.45s" repeatCount="indefinite" begin="-0.6000000000000001s"></animate></line></g><g><line x1="46%" y1="50%" x2="82%" y2="66%" stroke="#1C2225" stroke-width="1"></line><line x1="46%" y1="50%" x2="82%" y2="66%" stroke="#2faa6e" stroke-width="1" stroke-opacity="0.4" stroke-dasharray="40 80"><animate attributeName="stroke-dashoffset" from="120" to="0" dur="0.35s" repeatCount="indefinite" begin="-0.3s"></animate></line><line x1="46%" y1="50%" x2="82%" y2="66%" stroke="#2faa6e" stroke-width="1" stroke-opacity="0.25" stroke-dasharray="30 90"><animate attributeName="stroke-dashoffset" from="0" to="120" dur="0.35s" repeatCount="indef
1inite" begin="-0.7s"></animate></line></g><g><line x1="46%" y1="50%" x2="63%" y2="76%" stroke="#1C2225" stroke-width="1"></line><line x1="46%" y1="50%" x2="63%" y2="76%" stroke="#2faa6e" stroke-width="1" stroke-opacity="0.4" stroke-dasharray="40 80"><animate attributeName="stroke-dashoffset" from="120" to="0" dur="0.4s" repeatCount="indefinite" begin="-0.05s"></animate></line><line x1="46%" y1="50%" x2="63%" y2="76%" stroke="#2faa6e" stroke-width="1" stroke-opacity="0.25" stroke-dasharray="30 90"><animate attributeName="stroke-dashoffset" from="0" to="120" dur="0.4s" repeatCount="indefinite" begin="-0.45s"></animate></line></g><g><line x1="12%" y1="68%" x2="63%" y2="76%" stroke="#1C2225" stroke-width="1"></line><line x1="12%" y1="68%" x2="63%" y2="76%" stroke="#2faa6e" stroke-width="1" stroke-opacity="0.4" stroke-dasharray="40 80"><animate attributeName="stroke-dashoffset" from="120" to="0" dur="0.5s" repeatCount="indefinite" begin="-0.45s"></animate></line><line x1="12%" y1="68%" x2="63%" y2="76%" stroke="#2faa6e" stroke-width="1" stroke-opacity="0.25" stroke-dasharray="30 90"><animate attributeName="stroke-dashoffset" from="0" to="120" dur="0.5s" repeatCount="indefinite" begin="-0.8500000000000001s"></animate></line></g><g><line x1="82%" y1="66%" x2="63%" y2="76%" stroke="#1C2225" stroke-width="1"></line><line x1="82%" y1="66%" x2="63%" y2="76%" stroke="#2faa6e" stroke-width="1" stroke-opacity="0.4" stroke-dasharray="40 80"><animate attributeName="stroke-dashoffset" from="120" to="0" dur="0.45s" repeatCount="indefinite" begin="-0.35s"></animate></line><line x1="82%" y1="66%" x2="63%" y2="76%" stroke="#2faa6e" stroke-width="1" stroke-opacity="0.25" stroke-dasharray="30 90"><animate attributeName="stroke-dashoffset" from="0" to="120" dur="0.45s" repeatCount="indefinite" begin="-0.75s"></animate></line></g></svg><div class="absolute" style="left:10%;top:12%"><div class="border border-[#2D3335] bg-[#0d1416] p-3 relative z-10 w-[150px]"><div class="flex items-center justify-between mb-2"><span class="font-mono text-xs tracking-wider" style="color:#2faa6e">GPU-0</span><span class="w-1.5 h-1.5 rounded-full pulse-glow" style="background-color:#2faa6e"></span></div><div class="font-mono text-[10px] tracking-wider text-[#5B5F61] uppercase mb-1">GPU</div><div class="font-mono text-[10px] text-[#7e8385] mb-2">FLUX-[dev]</div><div class="h-1 bg-[#141B1D] w-full"><div class="h-full" style="width:94%;background-color:#2faa6e99"></div></div><div class="font-mono text-[10px] text-[#404647] mt-1">94<!-- -->% UTIL</div></div></div><div class="absolute" style="left:42%;top:5%"><div class="border border-[#2D3335] bg-[#0d1416] p-3 relative z-10 w-[150px]"><div class="flex items-center justify-between mb-2"><span class="font-mono text-xs tracking-wider" style="color:#3b82f6">ASIC-0</span><span class="w-1.5 h-1.5 rounded-full pulse-glow" style="background-color:#3b82f6"></span></div><div class="font-mono text-[10px] tracking-wider text-[#5B5F61] uppercase mb-1">ASIC</div><div class="font-mono text-[10px] text-[#7e8385] mb-2">FLUX-[dev]</div><div class="h-1 bg-[#141B1D] w-full"><div class="h-full" style="width:72%;background-color:#3b82f699"></div></div><div class="font-mono text-[10px] text-[#404647] mt-1">72<!-- -->% UTIL</div></div></div><div class="absolute" style="left:76%;top:14%"><div class="border border-[#2D3335] bg-[#0d1416] p-3 relative z-10 w-[150px]"><div class="flex items-center justify-between mb-2"><span class="font-mono text-xs tracking-wider" style="color:#2faa6e">GPU-1</span><span class="w-1.5 h-1.5 rounded-full pulse-glow" style="background-color:#2faa6e"></span></div><div class="font-mono text-[10px] tracking-wider text-[#5B5F61] uppercase mb-1">GPU</div><div class="font-mono text-[10px] text-[#7e8385] mb-2">Subagents 16-31</div><div class="h-1 bg-[#141B1D] w-full"><div class="h-full" style="width:87%;background-color:#2faa6e99"></div></div><div class="font-mono text-[10px] text-[#404647] mt-1">87<!-- -->% UTIL</div></div></div><div class="absolute" style="left:4%;top:60%"><div class="border border-[#2D3335] bg-[#0d1416] p-3 relative z-10 w-[150px]"><div class="flex items-center justify-between mb-2"><span class="font-mono text-xs tracking-wider" style="color:#a855f7">CPU-0</span><span class="w-1.5 h-1.5 rounded-full pulse-glow" style="background-color:#a855f7"></span></div><div class="font-mono text-[10px] tracking-wider text-[#5B5F61] uppercase mb-1">CPU</div><div class="font-mono text-[10px] text-[#7e8385] mb-2">Preprocessing</div><div class="h-1 bg-[#141B1D] w-full"><div class="h-full" style="width:45%;background-color:#a855f799"></div></div><div class="font-mono text-[10px] text-[#404647] mt-1">45<!-- -->% UTIL</div></div></div><div class="absolute" style="left:38%;top:42%"><div class="border border-[#2D3335] bg-[#0d1416] p-3 relative z-10 w-[150px]"><div class="flex items-center justify-between mb-2"><span class="font-mono text-xs tracking-wider" style="color:#2faa6e">GPU-2</span><span class="w-1.5 h-1.5 rounded-full pulse-glow" style="background-color:#2faa6e"></span></div><div class="font-mono text-[10px] tracking-wider text-[#5B5F61] uppercase mb-1">GPU</div><div class="font-mono text-[10px] text-[#7e8385] mb-2">Primary Agent LLM</div><div class="h-1 bg-[#141B1D] w-full"><div class="h-full" style="width:91%;background-color:#2faa6e99"></div></div><div class="font-mono text-[10px] text-[#404647] mt-1">91<!-- -->% UTIL</div></div></div><div class="absolute" style="left:74%;top:58%"><div class="border border-[#2D3335] bg-[#0d1416] p-3 relative z-10 w-[150px]"><div class="flex items-center justify-between mb-2"><span class="font-mono text-xs tracking-wider" style="color:#3b82f6">ASIC-1</span><span class="w-1.5 h-1.5 rounded-full pulse-glow" style="background-color:#3b82f6"></span></div><div class="font-mono text-[10px] tracking-wider text-[#5B5F61] uppercase mb-1">ASIC</div><div class="font-mono text-[10px] text-[#7e8385] mb-2">Subagents 0-15</div><div class="h-1 bg-[#141B1D] w-full"><div class="h-full" style="width:68%;background-color:#3b82f699"></div></div><div class="font-mono text-[10px] text-[#404647] mt-1">68<!-- -->% UTIL</div></div></div><div class="absolute" style="left:55%;top:68%"><div class="border border-[#2D3335] bg-[#0d1416] p-3 relative z-10 w-[150px]"><div class="flex items-center justify-between mb-2"><span class="font-mono text-xs tracking-wider" style="color:#2faa6e">GPU-3</span><span class="w-1.5 h-1.5 rounded-full pulse-glow" style="background-color:#2faa6e"></span></div><div class="font-mono text-[10px] tracking-wider text-[#5B5F61] uppercase mb-1">GPU</div><div class="font-mono text-[10px] text-[#7e8385] mb-2">Primary Agent LLM</div><div class="h-1 bg-[#141B1D] w-full"><div class="h-full" style="width:83%;background-color:#2faa6e99"></div></div><div class="font-mono text-[10px] text-[#404647] mt-1">83<!-- -->% UTIL</div></div></div></div><div class="md:hidden p-6"><div class="relative min-h-[420px]"><svg class="absolute inset-0 w-full h-full" preserveAspectRatio="none"><line x1="25%" y1="15%" x2="75%" y2="50%" stroke="#1C2225" stroke-width="1"></line><line x1="25%" y1="15%" x2="75%" y2="50%" stroke="#2faa6e" stroke-width="1" stroke-dasharray="40 80" stroke-dashoffset="120"><animate attributeName="stroke-dashoffset" from="120" to="0" dur="0.4s" begin="0s" repeatCount="indef
1inite"></animate></line><line x1="75%" y1="50%" x2="25%" y2="15%" stroke="#a855f7" stroke-width="1" stroke-dasharray="30 90" stroke-dashoffset="120"><animate attributeName="stroke-dashoffset" from="120" to="0" dur="0.45s" begin="-0.2s" repeatCount="indefinite"></animate></line><line x1="15%" y1="25%" x2="15%" y2="75%" stroke="#1C2225" stroke-width="1"></line><line x1="15%" y1="25%" x2="15%" y2="75%" stroke="#2faa6e" stroke-width="1" stroke-dasharray="40 80" stroke-dashoffset="120"><animate attributeName="stroke-dashoffset" from="120" to="0" dur="0.35s" begin="-0.15s" repeatCount="indefinite"></animate></line><line x1="15%" y1="75%" x2="15%" y2="25%" stroke="#3b82f6" stroke-width="1" stroke-dasharray="30 90" stroke-dashoffset="120"><animate attributeName="stroke-dashoffset" from="120" to="0" dur="0.4s" begin="-0.3s" repeatCount="indefinite"></animate></line><line x1="25%" y1="85%" x2="75%" y2="50%" stroke="#1C2225" stroke-width="1"></line><line x1="25%" y1="85%" x2="75%" y2="50%" stroke="#3b82f6" stroke-width="1" stroke-dasharray="40 80" stroke-dashoffset="120"><animate attributeName="stroke-dashoffset" from="120" to="0" dur="0.45s" begin="-0.1s" repeatCount="indefinite"></animate></line><line x1="75%" y1="50%" x2="25%" y2="85%" stroke="#a855f7" stroke-width="1" stroke-dasharray="30 90" stroke-dashoffset="120"><animate attributeName="stroke-dashoffset" from="120" to="0" dur="0.35s" begin="-0.25s" repeatCount="indefinite"></animate></line></svg><div class="absolute" style="left:2%;top:0%"><div class="border border-[#2D3335] bg-[#0d1416] p-3 relative z-10 w-[150px]"><div class="flex items-center justify-between mb-2"><span class="font-mono text-xs tracking-wider" style="color:#2faa6e">GPU-0</span><span class="w-1.5 h-1.5 rounded-full pulse-glow" style="background-color:#2faa6e"></span></div><div class="font-mono text-[10px] tracking-wider text-[#5B5F61] uppercase mb-1">GPU</div><div class="font-mono text-[10px] text-[#7e8385] mb-2">FLUX-[dev]</div><div class="h-1 bg-[#141B1D] w-full"><div class="h-full" style="width:94%;background-color:#2faa6e99"></div></div><div class="font-mono text-[10px] text-[#404647] mt-1">94<!-- -->% UTIL</div></div></div><div class="absolute" style="right:2%;top:50%;transform:translateY(-50%)"><div class="border border-[#2D3335] bg-[#0d1416] p-3 relative z-10 w-[150px]"><div class="flex items-center justify-between mb-2"><span class="font-mono text-xs tracking-wider" style="color:#a855f7">CPU-0</span><span class="w-1.5 h-1.5 rounded-full pulse-glow" style="background-color:#a855f7"></span></div><div class="font-mono text-[10px] tracking-wider text-[#5B5F61] uppercase mb-1">CPU</div><div class="font-mono text-[10px] text-[#7e8385] mb-2">Preprocessing</div><div class="h-1 bg-[#141B1D] w-full"><div class="h-full" style="width:45%;background-color:#a855f799"></div></div><div class="font-mono text-[10px] text-[#404647] mt-1">45<!-- -->% UTIL</div></div></div><div class="absolute" style="left:2%;bottom:0%"><div class="border border-[#2D3335] bg-[#0d1416] p-3 relative z-10 w-[150px]"><div class="flex items-center justify-between mb-2"><span class="font-mono text-xs tracking-wider" style="color:#3b82f6">ASIC-0</span><span class="w-1.5 h-1.5 rounded-full pulse-glow" style="background-color:#3b82f6"></span></div><div class="font-mono text-[10px] tracking-wider text-[#5B5F61] uppercase mb-1">ASIC</div><div class="font-mono text-[10px] text-[#7e8385] mb-2">FLUX-[dev]</div><div class="h-1 bg-[#141B1D] w-full"><div class="h-full" style="width:72%;background-color:#3b82f699"></div></div><div class="font-mono text-[10px] text-[#404647] mt-1">72<!-- -->% UTIL</div></div></div></div></div></div><div class="grid md:grid-cols-3 gap-0"><div class="p-6 border border-[#2D3335] "><div class="flex items-center justify-between mb-4"><h3 class="text-base font-medium text-[#D7D8D9]">Heterogeneous Compute</h3><span class="font-mono text-xs text-[#404647]">01</span></div><p class="text-sm text-[#7e8385] leading-relaxed">Inference across CPUs, GPUs, and ASICs deliver maximum throughput and superior TCO.</p></div><div class="p-6 border border-[#2D3335] md:border-l-0"><div class="flex items-center justify-between mb-4"><h3 class="text-base font-medium text-[#D7D8D9]">Dynamic Load Balancing</h3><span class="font-mono text-xs text-[#404647]">02</span></div><p class="text-sm text-[#7e8385] leading-relaxed">Continuously monitors utilization across every node and redistributes work in real time to eliminate bottlenecks and hotspots.</p></div><div class="p-6 border border-[#2D3335] md:border-l-0"><div class="flex items-center justify-between mb-4"><h3 class="text-base font-medium text-[#D7D8D9]">Lightning Quick Scaling</h3><span class="font-mono text-xs text-[#404647]">03</span></div><p class="text-sm text-[#7e8385] leading-relaxed">Nodes are dynamically booted and shutdown as workloads fluctuate, meeting peak loads without excess idle capacity.</p></div></div></div></div></div></section><section><div class="max-w-[1200px] mx-auto px-0 sm:px-6"><div class="sm:border-x border-[#2D3335] border-x-0"><div class="border-b border-[#2D3335] px-8 md:px-16 py-4"><span class="label-accent">Performance</span></div><div class="px-8 md:px-16 py-16"><div class="grid lg:grid-cols-2 gap-16"><div><h2 class="text-2xl sm:text-3xl font-medium text-[#D7D8D9] tracking-tight mb-6">Unmatched throughput</h2><p class="text-[#A1A4A5] leading-relaxed mb-10">Our compiler-first approach eliminates runtime overhead entirely. Models compiled by Luminal consistently outperform existing inference engines by 2-3x on standard benchmarks.</p><div class="grid grid-cols-3 border border-[#2D3335]"><div class="p-4 text-center "><div class="text-xl font-mono font-medium text-[#2faa6e] mb-1">3.2x</div><div class="label">vs vLLM</div></div><div class="p-4 text-center border-l border-[#2D3335]"><div class="text-xl font-mono font-medium text-[#2faa6e] mb-1"><10ms</div><div class="label">p99 latency</div></div><div class="p-4 text-center border-l border-[#2D3335]"><div class="text-xl font-mono font-medium text-[#2faa6e] mb-1">0%</div><div class="label">overhead</div></div></div></div><div class="border border-[#2D3335]"><div class="border-b border-[#2D3335] px-5 py-3"><span class="label">Tokens/sec, GPT-OSS 120B, 8xH100 SXM</span></div><div class="p-5 space-y-5"><div><div class="flex justify-between text-xs font-mono mb-1.5"><span class="text-[#2faa6e] tracking-wider">Luminal</span><span class="text-[#2faa6e]">36<!-- -->k tok/s</span></div><div class="h-2 bg-[#141B1D]"><div class="h-full bg-[#2faa6e]" style="width:100%"></div></div></div><div><div class="flex justify-between text-xs font-mono mb-1.5"><span class="text-[#5B5F61] tracking-wider">TensorRT-LLM</span><span class="text-[#404647]">28<!-- -->k tok/s</span></div><div class="h-2 bg-[#141B1D]"><div class="h-full bg-[#2D3335]" style="width:77.77777777777779%"></div></div></div><div><div class="flex justify-between text-xs font-mono mb-1.5"><span class="text-[#5B5F61] tracking-wider">vLLM</span><span class="text-[#404647]">26<!-- -->k tok/s</span></div><div class="h-2 bg-[#141B1D]"><div class="h-full bg-[#2D3335]" style="width:72.22222222222221%"></div></div></div><div><div class="flex justify-between text-xs font-mono mb-1.5"><span class="text-[#5B5F61] tracking-wider">PyTorch</span><span class="text-[#404647]">3<!-- -->k tok/s</span></div><div class="h-2 bg-[#141B1D]"><div class="h-full bg-[#2D3335]" style="width:8.333333333333332%"></div></div></div></div></div></div></div></div></div></section><section><div class="max-w-[1200px] mx-auto px-0 sm:px-6"><div class="sm:border-x border-[#2D3335] border-x-0"><div class="border-b border-[#2D3335] px-8 md:px-16 py-4"><span class="label-accent">Deployment</span></div><div id="waitlist" class="px-8 md:px-16 py-16"><h2 class="text-2xl sm:text-3xl font-medium text-[#D7D8D9] tracking-tight mb-16">Choose your deployment</h2><div class="grid md:grid-cols-2 gap-0"><div class="border border-[#2D3335] p-8 flex flex-col"><div class="flex items-center gap-3 mb-2"><span class="label-accent">Luminal Cloud</span></div><p class="text-sm text-[#7e8385] mb-8">Managed serverless inference. Deploy in minutes, scale automatically, and only pay for what you use.</p><ul class="space-y-3 mb-8"><li class="flex items-start gap-3 text-sm text-[#A1A4A5]">
1<svg class="w-3.5 h-3.5 text-[#2faa6e] shrink-0 mt-0.5" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2.5" stroke-linecap="round" stroke-linejoin="round"><polyline points="20 6 9 17 4 12"></polyline></svg>Serverless inference endpoints</li><li class="flex items-start gap-3 text-sm text-[#A1A4A5]"><svg class="w-3.5 h-3.5 text-[#2faa6e] shrink-0 mt-0.5" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2.5" stroke-linecap="round" stroke-linejoin="round"><polyline points="20 6 9 17 4 12"></polyline></svg>Scale to zero capabilities</li><li class="flex items-start gap-3 text-sm text-[#A1A4A5]"><svg class="w-3.5 h-3.5 text-[#2faa6e] shrink-0 mt-0.5" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2.5" stroke-linecap="round" stroke-linejoin="round"><polyline points="20 6 9 17 4 12"></polyline></svg>Automatic batching</li><li class="flex items-start gap-3 text-sm text-[#A1A4A5]"><svg class="w-3.5 h-3.5 text-[#2faa6e] shrink-0 mt-0.5" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2.5" stroke-linecap="round" stroke-linejoin="round"><polyline points="20 6 9 17 4 12"></polyline></svg>Optimized compilation</li><li class="flex items-start gap-3 text-sm text-[#A1A4A5]"><svg class="w-3.5 h-3.5 text-[#2faa6e] shrink-0 mt-0.5" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2.5" stroke-linecap="round" stroke-linejoin="round"><polyline points="20 6 9 17 4 12"></polyline></svg>Pay only for what you use</li></ul><button class="block w-full text-center bg-[#2faa6e] text-[#080F11] border border-[#2faa6e] py-2.5 text-sm font-medium font-mono uppercase tracking-wider hover:bg-transparent hover:text-[#2faa6e] transition-all duration-200 rounded-sm mt-auto cursor-pointer">Get Started</button></div><div class="border border-[#2D3335] md:border-l-0 p-8 flex flex-col"><div class="flex items-center gap-3 mb-2"><span class="label">On-Prem Deployment</span></div><p class="text-sm text-[#7e8385] mb-8">Run Luminal on your own infrastructure with dedicated support and enterprise-grade security.</p><ul class="space-y-3 mb-8"><li class="flex items-start gap-3 text-sm text-[#A1A4A5]"><svg class="w-3.5 h-3.5 text-[#2faa6e] shrink-0 mt-0.5" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2.5" stroke-linecap="round" stroke-linejoin="round"><polyline points="20 6 9 17 4 12"></polyline></svg>Licensed cloud or on-prem deployment</li><li class="flex items-start gap-3 text-sm text-[#A1A4A5]"><svg class="w-3.5 h-3.5 text-[#2faa6e] shrink-0 mt-0.5" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2.5" stroke-linecap="round" stroke-linejoin="round"><polyline points="20 6 9 17 4 12"></polyline></svg>Dedicated engineering support</li><li class="flex items-start gap-3 text-sm text-[#A1A4A5]"><svg class="w-3.5 h-3.5 text-[#2faa6e] shrink-0 mt-0.5" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2.5" stroke-linecap="round" stroke-linejoin="round"><polyline points="20 6 9 17 4 12"></polyline></svg>Custom kernel optimization</li><li class="flex items-start gap-3 text-sm text-[#A1A4A5]"><svg class="w-3.5 h-3.5 text-[#2faa6e] shrink-0 mt-0.5" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2.5" stroke-linecap="round" stroke-linejoin="round"><polyline points="20 6 9 17 4 12"></polyline></svg>Strict SLAs tailored to you</li></ul><button class="block w-full text-center border border-[#2D3335] text-[#A1A4A5] py-2.5 text-sm font-medium font-mono uppercase tracking-wider hover:border-[#404647] hover:text-[#D7D8D9] transition-colors rounded-sm mt-auto cursor-pointer">Talk to an Engineer</button></div></div></div></div></div></section><section><div class="max-w-[1200px] mx-auto px-0 sm:px-6"><div class="sm:border-x border-[#2D3335] border-x-0"><div class="border-b border-[#2D3335] px-8 md:px-16 py-4"><span class="label-accent">Get Started</span></div><div class="px-8 md:px-16 py-24 text-center"><h2 class="text-2xl sm:text-3xl font-medium text-[#D7D8D9] tracking-tight mb-4">Ready to accelerate your inference?</h2><p class="text-[#7e8385] mb-12 max-w-lg mx-auto">Get early access to the fastest AI inference platform in the world.</p><button class="bg-[#2faa6e] text-[#080F11] border border-[#2faa6e] px-8 py-2.5 text-sm font-medium font-mono uppercase tracking-wider hover:bg-transparent hover:text-[#2faa6e] transition-all duration-200 rounded-sm cursor-pointer">Get Started</button></div></div></div></section><footer><div class="max-w-[1200px] mx-auto px-0 sm:px-6"><div class="sm:border-x border-b border-[#2D3335]"><div class="px-8 md:px-16 py-6 flex flex-col sm:flex-row items-center justify-between gap-4"><span class="flex items-center gap-2 font-mono text-[#2faa6e] text-xs tracking-wider"><span class="relative w-3.5 h-3.5 rounded-full bg-[#1a5c3a] flex items-center justify-center"><span class="w-1.5 h-1.5 rounded-full bg-[#2faa6e]"></span></span>LUMINAL</span><div class="flex items-center gap-6"><div class="flex items-center gap-2 label"><div class="w-1.5 h-1.5 rounded-full bg-[#2faa6e] pulse-glow"></div>All Systems Online</div><button class="label hover:text-[#D7D8D9] transition-colors cursor-pointer">Contact Us</button></div><span class="label">© 2026 Luminal AI Inc.</span></div></div></div></footer></div><!--$--><!--/$-->
1<script src="/_next/static/chunks/8e44e41d890b5e5b.js?dpl=dpl_AWxxCfXD9UY7YUK42akeB8WWH9fr" id="_R_" async=""></script>
1<script>(self.__next_f=self.__next_f||[]).push([0])</script>
1<script>self.__next_f.push([1,"1:\"$Sreact.fragment\"\n2:I[39756,[\"/_next/static/chunks/ff1a16fafef87110.js?dpl=dpl_AWxxCfXD9UY7YUK42akeB8WWH9fr\",\"/_next/static/chunks/d2be314c3ece3fbe.js?dpl=dpl_AWxxCfXD9UY7YUK42akeB8WWH9fr\"],\"default\"]\n3:I[37457,[\"/_next/static/chunks/ff1a16fafef87110.js?dpl=dpl_AWxxCfXD9UY7YUK42akeB8WWH9fr\",\"/_next/static/chunks/d2be314c3ece3fbe.js?dpl=dpl_AWxxCfXD9UY7YUK42akeB8WWH9fr\"],\"default\"]\n4:I[47257,[\"/_next/static/chunks/ff1a16fafef87110.js?dpl=dpl_AWxxCfXD9UY7YUK42akeB8WWH9fr\",\"/_next/static/chunks/d2be314c3ece3fbe.js?dpl=dpl_AWxxCfXD9UY7YUK42akeB8WWH9fr\"],\"ClientPageRoot\"]\n5:I[52683,[\"/_next/static/chunks/fcb51f43b14e3807.js?dpl=dpl_AWxxCfXD9UY7YUK42akeB8WWH9fr\"],\"default\"]\n8:I[97367,[\"/_next/static/chunks/ff1a16fafef87110.js?dpl=dpl_AWxxCfXD9UY7YUK42akeB8WWH9fr\",\"/_next/static/chunks/d2be314c3ece3fbe.js?dpl=dpl_AWxxCfXD9UY7YUK42akeB8WWH9fr\"],\"OutletBoundary\"]\n9:\"$Sreact.suspense\"\nb:I[97367,[\"/_next/static/chunks/ff1a16fafef87110.js?dpl=dpl_AWxxCfXD9UY7YUK42akeB8WWH9fr\",\"/_next/static/chunks/d2be314c3ece3fbe.js?dpl=dpl_AWxxCfXD9UY7YUK42akeB8WWH9fr\"],\"ViewportBoundary\"]\nd:I[97367,[\"/_next/static/chunks/ff1a16fafef87110.js?dpl=dpl_AWxxCfXD9UY7YUK42akeB8WWH9fr\",\"/_next/static/chunks/d2be314c3ece3fbe.js?dpl=dpl_AWxxCfXD9UY7YUK42akeB8WWH9fr\"],\"MetadataBoundary\"]\nf:I[68027,[],\"default\"]\n:HL[\"/_next/static/chunks/29c48147fb5b609b.css?dpl=dpl_AWxxCfXD9UY7YUK42akeB8WWH9fr\",\"style\"]\n:HL[\"/_next/static/media/797e433ab948586e-s.p.29207c2f.woff2\",\"font\",{\"crossOrigin\":\"\",\"type\":\"font/woff2\"}]\n:HL[\"/_next/static/media/caa3a2e1cccd8315-s.p.3b6cae6d.woff2\",\"font\",{\"crossOrigin\":\"\",\"type\":\"font/woff2\"}]\n"])</script>
1<script>self.__next_f.push([1,"0:{\"P\":null,\"b\":\"RW1gn8FFwfPCFEPOXxt4k\",\"c\":[\"\",\"\"],\"q\":\"\",\"i\":false,\"f\":[[[\"\",{\"children\":[\"__PAGE__\",{}]},\"$undefined\",\"$undefined\",true],[[\"$\",\"$1\",\"c\",{\"children\":[[[\"$\",\"link\",\"0\",{\"rel\":\"stylesheet\",\"href\":\"/_next/static/chunks/29c48147fb5b609b.css?dpl=dpl_AWxxCfXD9UY7YUK42akeB8WWH9fr\",\"precedence\":\"next\",\"crossOrigin\":\"$undefined\",\"nonce\":\"$undefined\"}]],[\"$\",\"html\",null,{\"lang\":\"en\",\"className\":\"dark\",\"children\":[\"$\",\"body\",null,{\"className\":\"geist_a71539c9-module__T19VSG__variable geist_mono_8d43a2aa-module__8Li5zG__variable antialiased bg-[#030712] text-gray-100\",\"children\":[\"$\",\"$L2\",null,{\"parallelRouterKey\":\"children\",\"error\":\"$undefined\",\"errorStyles\":\"$undefined\",\"errorScripts\":\"$undefined\",\"template\":[\"$\",\"$L3\",null,{}],\"templateStyles\":\"$undefined\",\"templateScripts\":\"$undefined\",\"notFound\":[[[\"$\",\"title\",null,{\"children\":\"404: This page could not be found.\"}],[\"$\",\"div\",null,{\"style\":{\"fontFamily\":\"system-ui,\\\"Segoe UI\\\",Roboto,Helvetica,Arial,sans-serif,\\\"Apple Color Emoji\\\",\\\"Segoe UI Emoji\\\"\",\"height\":\"100vh\",\"textAlign\":\"center\",\"display\":\"flex\",\"flexDirection\":\"column\",\"alignItems\":\"center\",\"justifyContent\":\"center\"},\"children\":[\"$\",\"div\",null,{\"children\":[[\"$\",\"style\",null,{\"dangerouslySetInnerHTML\":{\"__html\":\"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}\"}}],[\"$\",\"h1\",null,{\"className\":\"next-error-h1\",\"style\":{\"display\":\"inline-block\",\"margin\":\"0 20px 0 0\",\"padding\":\"0 23px 0 0\",\"fontSize\":24,\"fontWeight\":500,\"verticalAlign\":\"top\",\"lineHeight\":\"49px\"},\"children\":404}],[\"$\",\"div\",null,{\"style\":{\"display\":\"inline-block\"},\"children\":[\"$\",\"h2\",null,{\"style\":{\"fontSize\":14,\"fontWeight\":400,\"lineHeight\":\"49px\",\"margin\":0},\"children\":\"This page could not be found.\"}]}]]}]}]],[]],\"forbidden\":\"$undefined\",\"unauthorized\":\"$undefined\"}]}]}]]}],{\"children\":[[\"$\",\"$1\",\"c\",{\"children\":[[\"$\",\"$L4\",null,{\"Component\":\"$5\",\"serverProvidedParams\":{\"searchParams\":{},\"params\":{},\"promises\":[\"$@6\",\"$@7\"]}}],[[\"$\",\"script\",\"script-0\",{\"src\":\"/_next/static/chunks/fcb51f43b14e3807.js?dpl=dpl_AWxxCfXD9UY7YUK42akeB8WWH9fr\",\"async\":true,\"nonce\":\"$undefined\"}]],[\"$\",\"$L8\",null,{\"children\":[\"$\",\"$9\",null,{\"name\":\"Next.MetadataOutlet\",\"children\":\"$@a\"}]}]]}],{},null,false,false]},null,false,false],[\"$\",\"$1\",\"h\",{\"children\":[null,[\"$\",\"$Lb\",null,{\"children\":\"$Lc\"}],[\"$\",\"div\",null,{\"hidden\":true,\"children\":[\"$\",\"$Ld\",null,{\"children\":[\"$\",\"$9\",null,{\"name\":\"Next.Metadata\",\"children\":\"$Le\"}]}]}],[\"$\",\"meta\",null,{\"name\":\"next-size-adjust\",\"content\":\"\"}]]}],false]],\"m\":\"$undefined\",\"G\":[\"$f\",[]],\"S\":true}\n"])</script>
1<script>self.__next_f.push([1,"6:{}\n7:\"$0:f:0:1:1:children:0:props:children:0:props:serverProvidedParams:params\"\n"])</script>
1<script>self.__next_f.push([1,"c:[[\"$\",\"meta\",\"0\",{\"charSet\":\"utf-8\"}],[\"$\",\"meta\",\"1\",{\"name\":\"viewport\",\"content\":\"width=device-width, initial-scale=1\"}]]\n"])</script>
1<script>self.__next_f.push([1,"10:I[27201,[\"/_next/static/chunks/ff1a16fafef87110.js?dpl=dpl_AWxxCfXD9UY7YUK42akeB8WWH9fr\",\"/_next/static/chunks/d2be314c3ece3fbe.js?dpl=dpl_AWxxCfXD9UY7YUK42akeB8WWH9fr\"],\"IconMark\"]\na:null\n"])</script>
1<script>self.__next_f.push([1,"e:[[\"$\",\"title\",\"0\",{\"children\":\"Luminal - Inference at the Speed of Light\"}],[\"$\",\"meta\",\"1\",{\"name\":\"description\",\"content\":\"Luminal is an AI inference compiler that compiles and optimizes AI models for GPUs and ASICs, delivering the fastest, highest throughput inference in the world.\"}],[\"$\",\"meta\",\"2\",{\"name\":\"keywords\",\"content\":\"Luminal,Luminal AI,AI inference,AI compiler,LLM inference,machine learning inference,model optimization,GPU inference,ASIC inference,AI inference engine,high throughput inference,low latency inference,AI model deployment,inference optimization,AI infrastructure\"}],[\"$\",\"meta\",\"3\",{\"name\":\"robots\",\"content\":\"index, follow\"}],[\"$\",\"link\",\"4\",{\"rel\":\"canonical\",\"href\":\"https://luminal.com\"}],[\"$\",\"meta\",\"5\",{\"property\":\"og:title\",\"content\":\"Luminal - Inference at the Speed of Light\"}],[\"$\",\"meta\",\"6\",{\"property\":\"og:description\",\"content\":\"Luminal is an AI inference compiler that compiles and optimizes AI models for GPUs and ASICs, delivering the fastest, highest throughput inference in the world.\"}],[\"$\",\"meta\",\"7\",{\"property\":\"og:url\",\"content\":\"https://luminal.com\"}],[\"$\",\"meta\",\"8\",{\"property\":\"og:site_name\",\"content\":\"Luminal\"}],[\"$\",\"meta\",\"9\",{\"property\":\"og:type\",\"content\":\"website\"}],[\"$\",\"meta\",\"10\",{\"name\":\"twitter:card\",\"content\":\"summary_large_image\"}],[\"$\",\"meta\",\"11\",{\"name\":\"twitter:title\",\"content\":\"Luminal - Inference at the Speed of Light\"}],[\"$\",\"meta\",\"12\",{\"name\":\"twitter:description\",\"content\":\"Luminal is an AI inference compiler that compiles and optimizes AI models for GPUs and ASICs, delivering the fastest, highest throughput inference in the world.\"}],[\"$\",\"link\",\"13\",{\"rel\":\"icon\",\"href\":\"/icon.svg?icon.92a21eff.svg\",\"sizes\":\"any\",\"type\":\"image/svg+xml\"}],[\"$\",\"$L10\",\"14\",{}]]\n"])</script>
1</body></html>
Line numbers count LF bytes from the start of the resource, as the search results do. Vendor segments are library code the classifier recognised; they are stored but not indexed. Bytes are shown as Latin1 characters, one per byte.