Spaces:
Running
Running
| <html lang="en"> | |
| <head> | |
| <meta charset="UTF-8"> | |
| <meta name="viewport" content="width=device-width, initial-scale=1.0"> | |
| <title>LLM Stats Benchmarks</title> | |
| <script src="https://cdn.tailwindcss.com"></script> | |
| <link href="https://fonts.googleapis.com/css2?family=Inter:wght@300;400;500;600;700&display=swap" rel="stylesheet"> | |
| <style> | |
| body { font-family: 'Inter', sans-serif; } | |
| .card-hover { transition: all 0.3s cubic-bezier(0.4, 0, 0.2, 1); } | |
| .card-hover:hover { transform: translateY(-4px); box-shadow: 0 20px 25px -5px rgba(0, 0, 0, 0.1), 0 10px 10px -5px rgba(0, 0, 0, 0.04); } | |
| .score-bar { transition: width 1s ease-out; } | |
| .fade-in { animation: fadeIn 0.5s ease-out; } | |
| @keyframes fadeIn { from { opacity: 0; transform: translateY(10px); } to { opacity: 1; transform: translateY(0); } } | |
| </style> | |
| </head> | |
| <body class="bg-gray-50 text-gray-900 min-h-screen flex flex-col"> | |
| <!-- Header --> | |
| <header class="bg-white border-b border-gray-200 sticky top-0 z-50 shadow-sm"> | |
| <div class="max-w-7xl mx-auto px-4 sm:px-6 lg:px-8 py-4"> | |
| <div class="flex flex-col md:flex-row md:items-center md:justify-between gap-4"> | |
| <div> | |
| <h1 class="text-2xl font-bold text-gray-900 flex items-center gap-2"> | |
| <svg class="w-8 h-8 text-indigo-600" fill="none" stroke="currentColor" viewBox="0 0 24 24"> | |
| <path stroke-linecap="round" stroke-linejoin="round" stroke-width="2" d="M9 19v-6a2 2 0 00-2-2H5a2 2 0 00-2 2v6a2 2 0 002 2h2a2 2 0 002-2zm0 0V9a2 2 0 012-2h2a2 2 0 012 2v10m-6 0a2 2 0 002 2h2a2 2 0 002-2m0 0V5a2 2 0 012-2h2a2 2 0 012 2v14a2 2 0 01-2 2h-2a2 2 0 01-2-2z"></path> | |
| </svg> | |
| LLM Stats Benchmarks | |
| </h1> | |
| <p class="text-sm text-gray-500 mt-1">Comprehensive leaderboard of AI model performance across diverse tasks</p> | |
| </div> | |
| <div class="flex flex-col sm:flex-row gap-3"> | |
| <input type="text" id="searchInput" placeholder="Search benchmarks..." | |
| class="px-4 py-2 border border-gray-300 rounded-lg focus:ring-2 focus:ring-indigo-500 focus:border-indigo-500 outline-none w-full sm:w-64 transition-shadow"> | |
| <select id="categoryFilter" | |
| class="px-4 py-2 border border-gray-300 rounded-lg focus:ring-2 focus:ring-indigo-500 focus:border-indigo-500 outline-none bg-white cursor-pointer"> | |
| <option value="all">All Categories</option> | |
| </select> | |
| </div> | |
| </div> | |
| </div> | |
| </header> | |
| <!-- Main Content --> | |
| <main class="flex-grow max-w-7xl mx-auto px-4 sm:px-6 lg:px-8 py-8 w-full"> | |
| <div id="benchmarkGrid" class="grid grid-cols-1 md:grid-cols-2 xl:grid-cols-3 gap-6"> | |
| <!-- Cards will be injected here --> | |
| </div> | |
| <div id="noResults" class="hidden text-center py-12 fade-in"> | |
| <svg class="w-16 h-16 text-gray-300 mx-auto mb-4" fill="none" stroke="currentColor" viewBox="0 0 24 24"> | |
| <path stroke-linecap="round" stroke-linejoin="round" stroke-width="2" d="M9.172 16.172a4 4 0 015.656 0M9 10h.01M15 10h.01M21 12a9 9 0 11-18 0 9 9 0 0118 0z"></path> | |
| </svg> | |
| <p class="text-gray-500 text-lg">No benchmarks found matching your criteria.</p> | |
| </div> | |
| </main> | |
| <!-- Footer --> | |
| <footer class="bg-white border-t border-gray-200 mt-auto"> | |
| <div class="max-w-7xl mx-auto px-4 sm:px-6 lg:px-8 py-6 text-center text-sm text-gray-500"> | |
| Data sourced from <a href="https://llm-stats.com/benchmarks" class="text-indigo-600 hover:underline font-medium" target="_blank" rel="noopener">llm-stats.com/benchmarks</a> | |
| <span class="mx-2">•</span> | |
| Built for Hugging Face Spaces | |
| </div> | |
| </footer> | |
| <script> | |
| const benchmarks = [ | |
| { | |
| name: "GPQA", | |
| category: "physics", | |
| description: "A challenging dataset of 448 multiple-choice questions written by domain experts in biology, physics, and chemistry. Questions are Google-proof and extremely difficult.", | |
| topModels: [ | |
| { rank: 1, name: "GPT-5.6 Sol", score: 94.6 }, | |
| { rank: 2, name: "Claude Mythos Preview", score: 94.6 }, | |
| { rank: 3, name: "Gemini 3.1 Pro", score: 94.3 }, | |
| { rank: 4, name: "Claude Opus 4.7", score: 94.2 }, | |
| { rank: 5, name: "Claude Opus 4.8", score: 93.6 } | |
| ] | |
| }, | |
| { | |
| name: "MMLU-Pro", | |
| category: "language", | |
| description: "A robust multi-task language understanding benchmark extending MMLU with 10 options, eliminating trivial questions, and focusing on reasoning-intensive tasks.", | |
| topModels: [ | |
| { rank: 1, name: "Qwen3.7 Max", score: 89.6 }, | |
| { rank: 2, name: "Qwen3.7-Plus", score: 88.5 }, | |
| { rank: 3, name: "Qwen3.6 Plus", score: 88.5 }, | |
| { rank: 4, name: "MiniMax M2.1", score: 88.0 }, | |
| { rank: 5, name: "Qwen3.5-397B-A17B", score: 87.8 } | |
| ] | |
| }, | |
| { | |
| name: "AIME 2025", | |
| category: "math", | |
| description: "All 30 problems from the 2025 American Invitational Mathematics Examination, testing olympiad-level mathematical reasoning with integer answers.", | |
| topModels: [ | |
| { rank: 1, name: "GPT-5.2 Pro", score: 100.0 }, | |
| { rank: 2, name: "GPT-5.2", score: 100.0 }, | |
| { rank: 3, name: "Gemini 3 Pro", score: 100.0 }, | |
| { rank: 4, name: "Kimi K2-Thinking-0905", score: 100.0 }, | |
| { rank: 5, name: "Grok-4 Heavy", score: 100.0 } | |
| ] | |
| }, | |
| { | |
| name: "SWE-Bench Verified", | |
| category: "reasoning", | |
| description: "A verified subset of 500 software engineering problems from real GitHub issues, evaluating language models' ability to resolve real-world coding issues.", | |
| topModels: [ | |
| { rank: 1, name: "Claude Fable 5", score: 95.0 }, | |
| { rank: 2, name: "Claude Mythos Preview", score: 93.9 }, | |
| { rank: 3, name: "Claude Opus 4.8", score: 88.6 }, | |
| { rank: 4, name: "Claude Opus 4.7", score: 87.6 }, | |
| { rank: 5, name: "Claude Sonnet 5", score: 85.2 } | |
| ] | |
| }, | |
| { | |
| name: "MMLU", | |
| category: "language", | |
| description: "Massive Multitask Language Understanding benchmark testing knowledge across 57 diverse subjects including STEM, humanities, and social sciences.", | |
| topModels: [ | |
| { rank: 1, name: "GPT-5", score: 92.5 }, | |
| { rank: 2, name: "o1", score: 91.8 }, | |
| { rank: 3, name: "GPT-4.5", score: 90.8 }, | |
| { rank: 4, name: "o1-preview", score: 90.8 }, | |
| { rank: 5, name: "Sarvam-105B", score: 90.6 } | |
| ] | |
| }, | |
| { | |
| name: "Humanity's Last Exam", | |
| category: "math", | |
| description: "A multi-modal academic benchmark with 2,500 questions across mathematics, humanities, and natural sciences, designed to test LLM capabilities at the frontier of human knowledge.", | |
| topModels: [ | |
| { rank: 1, name: "Claude Mythos Preview", score: 64.7 }, | |
| { rank: 2, name: "Claude Fable 5", score: 64.5 }, | |
| { rank: 3, name: "Muse Spark 1.1", score: 62.1 }, | |
| { rank: 4, name: "Muse Spark", score: 58.4 }, | |
| { rank: 5, name: "Claude Opus 4.8", score: 57.9 } | |
| ] | |
| }, | |
| { | |
| name: "LiveCodeBench", | |
| category: "reasoning", | |
| description: "A holistic and contamination-free evaluation benchmark for LLMs for code, continuously collecting new problems from programming contests (LeetCode, AtCoder, CodeForces).", | |
| topModels: [ | |
| { rank: 1, name: "DeepSeek-V4-Pro-Max", score: 93.5 }, | |
| { rank: 2, name: "DeepSeek-V4-Flash-Max", score: 91.6 }, | |
| { rank: 3, name: "DeepSeek-V3.2", score: 83.3 }, | |
| { rank: 4, name: "DeepSeek-V3.2 (Thinking)", score: 83.3 }, | |
| { rank: 5, name: "MiniMax M2", score: 83.0 } | |
| ] | |
| }, | |
| { | |
| name: "MATH", | |
| category: "math", | |
| description: "Contains 12,500 challenging competition mathematics problems from AMC 10, AMC 12, AIME, and other mathematics competitions with full step-by-step solutions.", | |
| topModels: [ | |
| { rank: 1, name: "o3-mini", score: 97.9 }, | |
| { rank: 2, name: "o1", score: 96.4 }, | |
| { rank: 3, name: "MiniStral 3 (14B)", score: 90.4 }, | |
| { rank: 4, name: "Mistral Large 3", score: 90.4 }, | |
| { rank: 5, name: "Gemini 2.0 Flash", score: 89.7 } | |
| ] | |
| }, | |
| { | |
| name: "HumanEval", | |
| category: "reasoning", | |
| description: "A benchmark that measures functional correctness for synthesizing programs from docstrings, consisting of 164 original programming problems.", | |
| topModels: [ | |
| { rank: 1, name: "MiniCPM-SALA", score: 95.1 }, | |
| { rank: 2, name: "Kimi K2 0905", score: 94.5 }, | |
| { rank: 3, name: "Claude 3.5 Sonnet", score: 93.7 }, | |
| { rank: 4, name: "GPT-5", score: 93.4 }, | |
| { rank: 5, name: "Kimi K2 Instruct", score: 93.3 } | |
| ] | |
| }, | |
| { | |
| name: "IFEval", | |
| category: "instruction following", | |
| description: "Instruction-Following Evaluation benchmark for large language models, focusing on verifiable instructions with 25 types of instructions and around 500 prompts.", | |
| topModels: [ | |
| { rank: 1, name: "Qwen3.5-27B", score: 95.0 }, | |
| { rank: 2, name: "Qwen3.7-Plus", score: 94.6 }, | |
| { rank: 3, name: "Qwen3.7 Max", score: 94.3 }, | |
| { rank: 4, name: "Qwen3.6 Plus", score: 94.3 }, | |
| { rank: 5, name: "o3-mini", score: 93.9 } | |
| ] | |
| }, | |
| { | |
| name: "MMMU-Pro", | |
| category: "multimodal", | |
| description: "A robust multi-discipline multimodal understanding benchmark that enhances MMMU through filtering text-only answerable questions and introducing vision-only input settings.", | |
| topModels: [ | |
| { rank: 1, name: "Gemini 3.5 Flash", score: 83.6 }, | |
| { rank: 2, name: "GPT-5.5", score: 83.2 }, | |
| { rank: 3, name: "GPT-5.6 Sol", score: 83.0 }, | |
| { rank: 4, name: "Seed 2.1 Pro", score: 82.7 }, | |
| { rank: 5, name: "Seed 2.1 Turbo", score: 82.2 } | |
| ] | |
| }, | |
| { | |
| name: "MMMU", | |
| category: "multimodal", | |
| description: "Massive Multi-discipline Multimodal Understanding benchmark designed to evaluate multimodal models on college-level subject knowledge and deliberate reasoning.", | |
| topModels: [ | |
| { rank: 1, name: "Qwen3.6 Plus", score: 86.0 }, | |
| { rank: 2, name: "GPT-5.1", score: 85.4 }, | |
| { rank: 3, name: "GPT-5.1 Instant", score: 85.4 }, | |
| { rank: 4, name: "GPT-5.1 Thinking", score: 85.4 }, | |
| { rank: 5, name: "GPT-5", score: 84.2 } | |
| ] | |
| }, | |
| { | |
| name: "BrowseComp", | |
| category: "reasoning", | |
| description: "A benchmark comprising 1,266 questions that challenge AI agents to persistently navigate the internet in search of hard-to-find, entangled information.", | |
| topModels: [ | |
| { rank: 1, name: "Kimi K3", score: 91.2 }, | |
| { rank: 2, name: "GPT-5.6 Sol", score: 90.4 }, | |
| { rank: 3, name: "GPT-5.5 Pro", score: 90.1 }, | |
| { rank: 4, name: "GPT-5.6 Terra", score: 87.5 }, | |
| { rank: 5, name: "Claude Mythos Preview", score: 86.9 } | |
| ] | |
| }, | |
| { | |
| name: "AIME 2024", | |
| category: "math", | |
| description: "American Invitational Mathematics Examination 2024, consisting of 30 challenging mathematical reasoning problems from AIME I and AIME II competitions.", | |
| topModels: [ | |
| { rank: 1, name: "Grok-3 Mini", score: 95.8 }, | |
| { rank: 2, name: "o4-mini", score: 93.4 }, | |
| { rank: 3, name: "LongCat-Flash-Thinking", score: 93.3 }, | |
| { rank: 4, name: "Grok-3", score: 93.3 }, | |
| { rank: 5, name: "Gemini 2.5 Pro", score: 92.0 } | |
| ] | |
| }, | |
| { | |
| name: "GSM8k", | |
| category: "math", | |
| description: "Grade School Math 8K, a dataset of 8.5K high-quality linguistically diverse grade school math word problems requiring multi-step reasoning.", | |
| topModels: [ | |
| { rank: 1, name: "MiMo-V2.5-Pro", score: 99.6 }, | |
| { rank: 2, name: "Kimi K2 Instruct", score: 97.3 }, | |
| { rank: 3, name: "o1", score: 97.1 }, | |
| { rank: 4, name: "GPT-4.5", score: 97.0 }, | |
| { rank: 5, name: "Llama 3.1 405B Instruct", score: 96.8 } | |
| ] | |
| } | |
| ]; | |
| const categoryColors = { | |
| "math": "bg-blue-100 text-blue-800 border-blue-200", | |
| "reasoning": "bg-purple-100 text-purple-800 border-purple-200", | |
| "language": "bg-green-100 text-green-800 border-green-200", | |
| "multimodal": "bg-pink-100 text-pink-800 border-pink-200", | |
| "physics": "bg-indigo-100 text-indigo-800 border-indigo-200", | |
| "instruction following": "bg-yellow-100 text-yellow-800 border-yellow-200", | |
| "coding": "bg-red-100 text-red-800 border-red-200", | |
| "long context": "bg-teal-100 text-teal-800 border-teal-200", | |
| "legal": "bg-gray-100 text-gray-800 border-gray-200", | |
| "spatial reasoning": "bg-orange-100 text-orange-800 border-orange-200", | |
| "general": "bg-slate-100 text-slate-800 border-slate-200", | |
| "image to text": "bg-cyan-100 text-cyan-800 border-cyan-200" | |
| }; | |
| const categoryDisplay = { | |
| "math": "Math", | |
| "reasoning": "Reasoning", | |
| "language": "Language", | |
| "multimodal": "Multimodal", | |
| "physics": "Physics", | |
| "instruction following": "Instruction Following", | |
| "coding": "Coding", | |
| "long context": "Long Context", | |
| "legal": "Legal", | |
| "spatial reasoning": "Spatial Reasoning", | |
| "general": "General", | |
| "image to text": "Image to Text" | |
| }; | |
| // Populate category filter | |
| const categoryFilter = document.getElementById('categoryFilter'); | |
| const uniqueCategories = [...new Set(benchmarks.map(b => b.category))].sort(); | |
| uniqueCategories.forEach(cat => { | |
| const option = document.createElement('option'); | |
| option.value = cat; | |
| option.textContent = categoryDisplay[cat] || cat.charAt(0).toUpperCase() + cat.slice(1); | |
| categoryFilter.appendChild(option); | |
| }); | |
| function renderBenchmarks(data) { | |
| const grid = document.getElementById('benchmarkGrid'); | |
| const noResults = document.getElementById('noResults'); | |
| grid.innerHTML = ''; | |
| if (data.length === 0) { | |
| noResults.classList.remove('hidden'); | |
| return; | |
| } | |
| noResults.classList.add('hidden'); | |
| data.forEach(benchmark => { | |
| const colorClass = categoryColors[benchmark.category] || "bg-gray-100 text-gray-800 border-gray-200"; | |
| const categoryName = categoryDisplay[benchmark.category] || benchmark.category.charAt(0).toUpperCase() + benchmark.category.slice(1); | |
| const maxScore = Math.max(...benchmark.topModels.map(m => m.score)); | |
| const card = document.createElement('div'); | |
| card.className = 'bg-white rounded-xl border border-gray-200 p-6 card-hover fade-in flex flex-col'; | |
| card.innerHTML = ` | |
| <div class="flex items-start justify-between mb-3"> | |
| <span class="inline-flex items-center px-2.5 py-0.5 rounded-full text-xs font-medium border ${colorClass}"> | |
| ${categoryName} | |
| </span> | |
| </div> | |
| <h3 class="text-xl font-bold text-gray-900 mb-2">${benchmark.name}</h3> | |
| <p class="text-sm text-gray-600 mb-6 flex-grow leading-relaxed">${benchmark.description}</p> | |
| <div class="space-y-3 mt-auto"> | |
| <h4 class="text-xs font-semibold text-gray-500 uppercase tracking-wider">Top Performers</h4> | |
| ${benchmark.topModels.map((model, index) => ` | |
| <div class="flex items-center gap-3"> | |
| <span class="flex-shrink-0 w-5 h-5 flex items-center justify-center rounded-full text-xs font-bold ${index === 0 ? 'bg-yellow-100 text-yellow-700' : 'bg-gray-100 text-gray-600'}"> | |
| ${model.rank} | |
| </span> | |
| <div class="flex-grow min-w-0"> | |
| <div class="flex justify-between items-center mb-1"> | |
| <span class="text-sm font-medium text-gray-900 truncate">${model.name}</span> | |
| <span class="text-sm font-bold text-indigo-600">${model.score}%</span> | |
| </div> | |
| <div class="w-full bg-gray-100 rounded-full h-1.5"> | |
| <div class="score-bar bg-indigo-500 h-1.5 rounded-full" style="width: ${(model.score / 100) * 100}%"></div> | |
| </div> | |
| </div> | |
| </div> | |
| `).join('')} | |
| </div> | |
| `; | |
| grid.appendChild(card); | |
| }); | |
| } | |
| function filterBenchmarks() { | |
| const searchTerm = document.getElementById('searchInput').value.toLowerCase(); | |
| const selectedCategory = document.getElementById('categoryFilter').value; | |
| const filtered = benchmarks.filter(benchmark => { | |
| const matchesSearch = benchmark.name.toLowerCase().includes(searchTerm) || | |
| benchmark.description.toLowerCase().includes(searchTerm) || | |
| benchmark.topModels.some(m => m.name.toLowerCase().includes(searchTerm)); | |
| const matchesCategory = selectedCategory === 'all' || benchmark.category === selectedCategory; | |
| return matchesSearch && matchesCategory; | |
| }); | |
| renderBenchmarks(filtered); | |
| } | |
| document.getElementById('searchInput').addEventListener('input', filterBenchmarks); | |
| document.getElementById('categoryFilter').addEventListener('change', filterBenchmarks); | |
| // Initial render | |
| renderBenchmarks(benchmarks); | |
| </script> | |
| </body> | |
| </html> |