{
 "schema_version": 1,
 "published": true,
 "generated_at": "2026-11-17T05:51:19Z",
 "run_id": "20261117T055119Z-8ac220",
 "notice": "Scores are reproduced only from the listed sources, with attribution and as-of dates. Nothing is estimated. Compare models only within one leaderboard.",
 "categories": [
  "Aggregators",
  "General intelligence & reasoning",
  "Math",
  "Coding / software engineering",
  "Agentic, computer use & long-horizon",
  "Cybersecurity",
  "Design, UI & multimodal",
  "Knowledge, factuality & hallucination",
  "Long context",
  "Instruction following",
  "Safety, alignment & robustness",
  "Economic & real-world work value",
  "Multilingual (Swedish/Nordic)",
  "Other"
 ],
 "sources": [
  {
   "id": "epoch",
   "name": "Epoch AI Benchmarking Hub",
   "url": "https://epoch.ai/benchmarks",
   "data_url": "https://epoch.ai/data/benchmark_data.zip",
   "license": "CC BY 4.0",
   "status": "ok",
   "last_ok": "2026-11-17T05:51:19Z",
   "age_days": 0,
   "strategy": "zip",
   "boards": 13
  },
  {
   "id": "lmarena",
   "name": "Arena (LMArena) leaderboard dataset",
   "url": "https://lmarena.ai/leaderboard",
   "data_url": "https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset",
   "license": "CC BY 4.0",
   "status": "ok",
   "last_ok": "2026-11-17T05:51:19Z",
   "age_days": 0,
   "strategy": "rows-api",
   "boards": 3
  },
  {
   "id": "swebench",
   "name": "SWE-bench leaderboards",
   "url": "https://www.swebench.com/",
   "data_url": "https://github.com/SWE-bench/swe-bench.github.io",
   "license": "public repository data",
   "status": "ok",
   "last_ok": "2026-11-17T05:51:19Z",
   "age_days": 0,
   "strategy": "github-raw",
   "boards": 2
  },
  {
   "id": "euroeval",
   "name": "EuroEval Swedish leaderboard",
   "url": "https://euroeval.com/leaderboards/swedish",
   "data_url": "https://euroeval.com/leaderboards/swedish",
   "license": "EuroEval (MIT-licensed project); results credited to EuroEval",
   "status": "ok",
   "last_ok": "2026-11-17T05:51:19Z",
   "age_days": 0,
   "strategy": "generative-csv",
   "boards": 1
  }
 ],
 "leaderboards": [
  {
   "id": "epoch-eci",
   "title": "Epoch Capabilities Index (ECI)",
   "hint": "Epoch Capabilities Index",
   "category": "Aggregators",
   "unit": "ECI",
   "kind": "index",
   "higher_is_better": true,
   "as_of": "2026-11-17",
   "source_url": "https://epoch.ai/eci",
   "data_url": "https://epoch.ai/data/benchmark_data.zip",
   "attribution": "Epoch AI, Benchmarking Hub (CC BY 4.0)",
   "rows": [
    {
     "model": "Claude Fable 5.2",
     "org": "Anthropic",
     "score": 173.44,
     "rank": 1,
     "display": "173.4"
    },
    {
     "model": "Gemini 4 Carbon",
     "org": "Google DeepMind",
     "score": 169.76,
     "rank": 2,
     "display": "169.8"
    },
    {
     "model": "GPT-6 Astra",
     "org": "OpenAI",
     "score": 169.0,
     "rank": 3,
     "display": "169.0"
    },
    {
     "model": "Claude Opus 5.5",
     "org": "Anthropic",
     "score": 168.54,
     "rank": 4,
     "display": "168.5"
    },
    {
     "model": "GPT-6.1 Sol",
     "org": "OpenAI",
     "score": 168.48,
     "rank": 5,
     "display": "168.5"
    }
   ],
   "note": "IRT-based index across many benchmarks (higher is better).",
   "source_id": "epoch",
   "source_name": "Epoch AI Benchmarking Hub",
   "fetched_on": "2026-11-17",
   "stale": false,
   "age_days": 0,
   "benchmark_id": "epoch-capabilities-index-eci-benchmarking-hub",
   "featured": true
  },
  {
   "id": "arena-text",
   "title": "Arena Text (overall)",
   "hint": "Arena (formerly LMArena)",
   "category": "Aggregators",
   "unit": "Arena score",
   "kind": "elo",
   "higher_is_better": true,
   "as_of": "2026-11-11",
   "source_url": "https://lmarena.ai/leaderboard/text",
   "data_url": "https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset",
   "attribution": "Arena / LMArena leaderboard-dataset (CC BY 4.0)",
   "rows": [
    {
     "model": "Gemini 4 Carbon (High)",
     "org": "google",
     "score": 1538.4,
     "votes": 4273,
     "rank": 1,
     "display": "1538"
    },
    {
     "model": "Gemini 4 Argon (High)",
     "org": "google",
     "score": 1529.5,
     "votes": 20779,
     "rank": 2,
     "display": "1530"
    },
    {
     "model": "Claude Opus 5.5 (High)",
     "org": "anthropic",
     "score": 1513.7,
     "votes": 15793,
     "rank": 3,
     "display": "1514"
    },
    {
     "model": "Claude Fable 5.1 (Max)",
     "org": "anthropic",
     "score": 1512.0,
     "votes": 20511,
     "rank": 4,
     "display": "1512"
    },
    {
     "model": "Claude Opus 5 (Max)",
     "org": "anthropic",
     "score": 1506.5,
     "votes": 35670,
     "rank": 5,
     "display": "1506"
    }
   ],
   "note": "Crowd-voted human preference rating (Elo-style).",
   "source_id": "lmarena",
   "source_name": "Arena (LMArena) leaderboard dataset",
   "fetched_on": "2026-11-17",
   "stale": false,
   "age_days": 0,
   "benchmark_id": "arena-formerly-lmarena-text-webdev-vision",
   "featured": true
  },
  {
   "id": "epoch-hle",
   "title": "Humanity's Last Exam",
   "hint": "Humanity's Last Exam",
   "category": "General intelligence & reasoning",
   "unit": "%",
   "kind": "percent",
   "higher_is_better": true,
   "as_of": "2026-11-17",
   "source_url": "https://epoch.ai/benchmarks",
   "data_url": "https://epoch.ai/data/benchmark_data.zip",
   "attribution": "Epoch AI, Benchmarking Hub (CC BY 4.0)",
   "rows": [
    {
     "model": "GPT-6 Astra",
     "org": "OpenAI",
     "score": 54.8,
     "rank": 1,
     "display": "54.8%"
    },
    {
     "model": "Claude Fable 5.1 (xhigh)",
     "org": "Anthropic",
     "score": 46.5,
     "rank": 2,
     "display": "46.5%"
    },
    {
     "model": "Gemini 3.1 Pro Preview",
     "org": "Google DeepMind",
     "score": 46.4,
     "rank": 3,
     "display": "46.4%"
    },
    {
     "model": "Claude Sonnet 5.5 (xhigh)",
     "org": "Anthropic",
     "score": 46.1,
     "rank": 4,
     "display": "46.1%"
    },
    {
     "model": "Gemini 3.8 Flash",
     "org": "Google DeepMind",
     "score": 44.5,
     "rank": 5,
     "display": "44.5%"
    }
   ],
   "note": "Best result per model family.",
   "column": "Accuracy",
   "newest_entry": "2026-10-07",
   "source_id": "epoch",
   "source_name": "Epoch AI Benchmarking Hub",
   "fetched_on": "2026-11-17",
   "stale": false,
   "age_days": 0,
   "benchmark_id": "humanity-s-last-exam-hle-hle-rolling",
   "featured": true
  },
  {
   "id": "epoch-arc-agi-2",
   "title": "ARC-AGI-2",
   "hint": "ARC-AGI-3",
   "category": "General intelligence & reasoning",
   "unit": "%",
   "kind": "percent",
   "higher_is_better": true,
   "as_of": "2026-11-17",
   "source_url": "https://epoch.ai/benchmarks",
   "data_url": "https://epoch.ai/data/benchmark_data.zip",
   "attribution": "Epoch AI, Benchmarking Hub (CC BY 4.0)",
   "rows": [
    {
     "model": "GPT-6 Astra (max)",
     "org": "OpenAI",
     "score": 95.0,
     "rank": 1,
     "display": "95.0%"
    },
    {
     "model": "GPT-6.1 Sol (max)",
     "org": "OpenAI",
     "score": 94.2,
     "rank": 2,
     "display": "94.2%"
    },
    {
     "model": "Gemini 4 Carbon (medium)",
     "org": "Google DeepMind",
     "score": 94.2,
     "rank": 3,
     "display": "94.2%"
    },
    {
     "model": "Claude Fable 5.2 (xhigh)",
     "org": "Anthropic",
     "score": 93.8,
     "rank": 4,
     "display": "93.8%"
    },
    {
     "model": "Claude Opus 5.5 (high)",
     "org": "Anthropic",
     "score": 93.3,
     "rank": 5,
     "display": "93.3%"
    }
   ],
   "note": "Best result per model family.",
   "column": "Score",
   "newest_entry": "2026-11-10",
   "source_id": "epoch",
   "source_name": "Epoch AI Benchmarking Hub",
   "fetched_on": "2026-11-17",
   "stale": false,
   "age_days": 0,
   "benchmark_id": "arc-agi-3-with-arc-agi-2",
   "featured": true
  },
  {
   "id": "epoch-frontiermath-t4",
   "title": "FrontierMath Tier 4 (v2)",
   "hint": "FrontierMath Tiers 1–4",
   "category": "Math",
   "unit": "%",
   "kind": "percent",
   "higher_is_better": true,
   "as_of": "2026-11-17",
   "source_url": "https://epoch.ai/benchmarks/frontiermath-tier-4-v2",
   "data_url": "https://epoch.ai/data/benchmark_data.zip",
   "attribution": "Epoch AI, Benchmarking Hub (CC BY 4.0)",
   "rows": [
    {
     "model": "Claude Fable 5.2 (max)",
     "org": "Anthropic",
     "score": 100.0,
     "rank": 1,
     "display": "100.0%"
    },
    {
     "model": "GPT-6.1 Sol (max)",
     "org": "OpenAI",
     "score": 100.0,
     "rank": 2,
     "display": "100.0%"
    },
    {
     "model": "GPT-6 Astra (medium)",
     "org": "OpenAI",
     "score": 97.6,
     "rank": 3,
     "display": "97.6%"
    },
    {
     "model": "Gemini 4 Carbon (high)",
     "org": "Google DeepMind",
     "score": 95.1,
     "rank": 4,
     "display": "95.1%"
    },
    {
     "model": "Claude Opus 5.5 (max)",
     "org": "Anthropic",
     "score": 95.0,
     "rank": 5,
     "display": "95.0%"
    }
   ],
   "note": "Best result per model family.",
   "column": "Best score (across scorers)",
   "newest_entry": "2026-11-06",
   "source_id": "epoch",
   "source_name": "Epoch AI Benchmarking Hub",
   "fetched_on": "2026-11-17",
   "stale": false,
   "age_days": 0,
   "benchmark_id": "frontiermath-tiers-1-4-v2",
   "featured": true
  },
  {
   "id": "epoch-terminal-bench",
   "title": "Terminal-Bench 2.0",
   "hint": "Terminal-Bench",
   "category": "Coding / software engineering",
   "unit": "%",
   "kind": "percent",
   "higher_is_better": true,
   "as_of": "2026-11-17",
   "source_url": "https://epoch.ai/benchmarks",
   "data_url": "https://epoch.ai/data/benchmark_data.zip",
   "attribution": "Epoch AI, Benchmarking Hub (CC BY 4.0)",
   "rows": [
    {
     "model": "GPT-5.5",
     "org": "OpenAI",
     "score": 84.7,
     "agent": "NexAU-AHE",
     "rank": 1,
     "display": "84.7%"
    },
    {
     "model": "GPT-5.4",
     "org": "OpenAI",
     "score": 81.8,
     "agent": "ForgeCode",
     "rank": 2,
     "display": "81.8%"
    },
    {
     "model": "Gemini 3.1 Pro Preview",
     "org": "Google DeepMind",
     "score": 80.2,
     "agent": "TongAgents",
     "rank": 3,
     "display": "80.2%"
    },
    {
     "model": "Claude Opus 4.7",
     "org": "Anthropic",
     "score": 80.2,
     "agent": "WOZCODE",
     "rank": 4,
     "display": "80.2%"
    },
    {
     "model": "Claude Opus 4.6",
     "org": "Anthropic",
     "score": 79.8,
     "agent": "ForgeCode",
     "rank": 5,
     "display": "79.8%"
    }
   ],
   "note": "Best result per model family; one row per model (best agent shown).",
   "column": "Accuracy mean",
   "newest_entry": "2026-05-14",
   "source_id": "epoch",
   "source_name": "Epoch AI Benchmarking Hub",
   "fetched_on": "2026-11-17",
   "stale": false,
   "age_days": 0,
   "benchmark_id": "terminal-bench-4-0-and-3-0",
   "featured": true
  },
  {
   "id": "epoch-osworld-2",
   "title": "OSWorld 2.0",
   "hint": "OSWorld 2.0",
   "category": "Agentic, computer use & long-horizon",
   "unit": "%",
   "kind": "percent",
   "higher_is_better": true,
   "as_of": "2026-11-17",
   "source_url": "https://epoch.ai/benchmarks",
   "data_url": "https://epoch.ai/data/benchmark_data.zip",
   "attribution": "Epoch AI, Benchmarking Hub (CC BY 4.0)",
   "rows": [
    {
     "model": "Claude Opus 5 (max)",
     "org": "Anthropic",
     "score": 31.4,
     "rank": 1,
     "display": "31.4%"
    },
    {
     "model": "GPT-5.6 Sol (max)",
     "org": "OpenAI",
     "score": 27.3,
     "rank": 2,
     "display": "27.3%"
    },
    {
     "model": "Claude Opus 4.8",
     "org": "Anthropic",
     "score": 20.6,
     "rank": 3,
     "display": "20.6%"
    },
    {
     "model": "Claude Opus 4.7 (max)",
     "org": "Anthropic",
     "score": 18.2,
     "rank": 4,
     "display": "18.2%"
    },
    {
     "model": "GPT-5.5 (xhigh)",
     "org": "OpenAI",
     "score": 13.0,
     "rank": 5,
     "display": "13.0%"
    }
   ],
   "note": "Best result per model family.",
   "column": "Binary accuracy",
   "newest_entry": "2026-07-24",
   "source_id": "epoch",
   "source_name": "Epoch AI Benchmarking Hub",
   "fetched_on": "2026-11-17",
   "stale": false,
   "age_days": 0,
   "benchmark_id": "osworld-2-0-and-osworld-verified",
   "featured": true
  },
  {
   "id": "epoch-metr",
   "title": "METR time horizon (50%)",
   "hint": "METR Time Horizon",
   "category": "Agentic, computer use & long-horizon",
   "unit": "min",
   "kind": "minutes",
   "higher_is_better": true,
   "as_of": "2026-11-17",
   "source_url": "https://epoch.ai/benchmarks",
   "data_url": "https://epoch.ai/data/benchmark_data.zip",
   "attribution": "Epoch AI, Benchmarking Hub (CC BY 4.0)",
   "rows": [
    {
     "model": "Claude Mythos Preview (Early)",
     "org": "Anthropic",
     "score": 1044.8,
     "rank": 1,
     "display": "1045 min"
    },
    {
     "model": "Claude Opus 4.6",
     "org": "Anthropic",
     "score": 718.8,
     "rank": 2,
     "display": "719 min"
    },
    {
     "model": "Gemini 3.1 Pro Preview",
     "org": "Google DeepMind",
     "score": 384.1,
     "rank": 3,
     "display": "384 min"
    },
    {
     "model": "GPT-5.2 (high)",
     "org": "OpenAI",
     "score": 352.2,
     "rank": 4,
     "display": "352 min"
    },
    {
     "model": "GPT-5.3 Codex",
     "org": "OpenAI",
     "score": 349.5,
     "rank": 5,
     "display": "350 min"
    }
   ],
   "note": "Best result per model family.",
   "column": "Time horizon",
   "newest_entry": "2026-04-07",
   "source_id": "epoch",
   "source_name": "Epoch AI Benchmarking Hub",
   "fetched_on": "2026-11-17",
   "stale": false,
   "age_days": 0,
   "benchmark_id": "metr-time-horizon-1-1",
   "featured": true
  },
  {
   "id": "euroeval-swedish",
   "title": "EuroEval Swedish (generative)",
   "hint": "EuroEval",
   "category": "Multilingual (Swedish/Nordic)",
   "unit": "rank score",
   "kind": "rank-score",
   "higher_is_better": false,
   "as_of": "2026-11-15",
   "source_url": "https://euroeval.com/leaderboards/swedish",
   "data_url": "https://euroeval.com/leaderboards/swedish",
   "attribution": "EuroEval (euroeval.com)",
   "rows": [
    {
     "model": "gpt-6.1-sol (zero-shot, val)",
     "score": 1.22,
     "rank": 1,
     "display": "1.22"
    },
    {
     "model": "gpt-6-astra (zero-shot, val)",
     "score": 1.24,
     "rank": 2,
     "display": "1.24"
    },
    {
     "model": "gemini/gemini-3.7-flash (zero-shot, val)",
     "score": 1.28,
     "rank": 3,
     "display": "1.28"
    },
    {
     "model": "zai-org/GLM-5.3-Flash (val)",
     "score": 1.3,
     "rank": 4,
     "display": "1.30"
    },
    {
     "model": "gemini/gemini-3.6-flash (zero-shot, val)",
     "score": 1.33,
     "rank": 5,
     "display": "1.33"
    }
   ],
   "note": "EuroEval rank score across Swedish tasks: lower is better.",
   "source_id": "euroeval",
   "source_name": "EuroEval Swedish leaderboard",
   "fetched_on": "2026-11-17",
   "stale": false,
   "age_days": 0,
   "benchmark_id": "euroeval-formerly-scandeval-swedish-scandinavian",
   "featured": true
  },
  {
   "id": "arena-webdev",
   "title": "Arena WebDev",
   "hint": "Arena WebDev",
   "category": "Coding / software engineering",
   "unit": "Arena score",
   "kind": "elo",
   "higher_is_better": true,
   "as_of": "2026-11-15",
   "source_url": "https://lmarena.ai/leaderboard/webdev",
   "data_url": "https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset",
   "attribution": "Arena / LMArena leaderboard-dataset (CC BY 4.0)",
   "rows": [
    {
     "model": "Claude Fable 5.2 (xHigh)",
     "org": "anthropic",
     "score": 1826.0,
     "votes": 1445,
     "rank": 1,
     "display": "1826"
    },
    {
     "model": "Claude Opus 5.5 (Max)",
     "org": "anthropic",
     "score": 1815.6,
     "votes": 4399,
     "rank": 2,
     "display": "1816"
    },
    {
     "model": "Claude Sonnet 5.5 (xHigh)",
     "org": "anthropic",
     "score": 1772.2,
     "votes": 6120,
     "rank": 3,
     "display": "1772"
    },
    {
     "model": "Gemini 4 Carbon (High)",
     "org": "google",
     "score": 1768.3,
     "votes": 2105,
     "rank": 4,
     "display": "1768"
    },
    {
     "model": "GPT-6 Astra (Max)",
     "org": "openai",
     "score": 1762.9,
     "votes": 8685,
     "rank": 5,
     "display": "1763"
    }
   ],
   "note": "Crowd-voted human preference rating (Elo-style).",
   "source_id": "lmarena",
   "source_name": "Arena (LMArena) leaderboard dataset",
   "fetched_on": "2026-11-17",
   "stale": false,
   "age_days": 0,
   "benchmark_id": "arena-webdev",
   "featured": false
  },
  {
   "id": "swebench-multilingual",
   "title": "SWE-bench Multilingual",
   "hint": "SWE-bench Verified",
   "category": "Coding / software engineering",
   "unit": "% resolved",
   "kind": "percent",
   "higher_is_better": true,
   "as_of": "2026-02-20",
   "source_url": "https://www.swebench.com/",
   "data_url": "https://github.com/SWE-bench/swe-bench.github.io",
   "attribution": "SWE-bench team (swebench.com)",
   "rows": [
    {
     "model": "Gemini 3 Flash",
     "org": "SWE-agent",
     "score": 72.7,
     "date": "2026-02-13",
     "rank": 1,
     "display": "72.7%"
    },
    {
     "model": "Claude 4.6 Opus",
     "org": "SWE-agent",
     "score": 72.0,
     "date": "2026-02-13",
     "rank": 2,
     "display": "72.0%"
    },
    {
     "model": "Claude 4.5 Opus",
     "org": "SWE-agent",
     "score": 70.7,
     "date": "2026-02-13",
     "rank": 3,
     "display": "70.7%"
    },
    {
     "model": "GLM 5",
     "org": "SWE-agent",
     "score": 69.7,
     "date": "2026-02-13",
     "rank": 4,
     "display": "69.7%"
    },
    {
     "model": "Gemini 3 Pro",
     "org": "SWE-agent",
     "score": 68.7,
     "date": "2026-02-13",
     "rank": 5,
     "display": "68.7%"
    }
   ],
   "note": "Agent + model submissions; date = newest submission on the board.",
   "source_id": "swebench",
   "source_name": "SWE-bench leaderboards",
   "fetched_on": "2026-11-17",
   "stale": false,
   "age_days": 0,
   "benchmark_id": "swe-bench-verified-multimodal",
   "featured": false
  },
  {
   "id": "epoch-swe-bench-verified",
   "title": "SWE-bench Verified (Epoch runs)",
   "hint": "SWE-bench Verified",
   "category": "Coding / software engineering",
   "unit": "%",
   "kind": "percent",
   "higher_is_better": true,
   "as_of": "2026-11-17",
   "source_url": "https://epoch.ai/benchmarks",
   "data_url": "https://epoch.ai/data/benchmark_data.zip",
   "attribution": "Epoch AI, Benchmarking Hub (CC BY 4.0)",
   "rows": [
    {
     "model": "Claude Opus 4.7 (max)",
     "org": "Anthropic",
     "score": 83.5,
     "rank": 1,
     "display": "83.5%"
    },
    {
     "model": "GPT-5.5 (xhigh)",
     "org": "OpenAI",
     "score": 80.6,
     "rank": 2,
     "display": "80.6%"
    },
    {
     "model": "Gemini 3.5 Flash",
     "org": "Google DeepMind",
     "score": 79.3,
     "rank": 3,
     "display": "79.3%"
    },
    {
     "model": "GLM-5.2",
     "org": "Z.ai (Zhipu AI)",
     "score": 78.7,
     "rank": 4,
     "display": "78.7%"
    },
    {
     "model": "Claude Opus 4.6 (no thinking)",
     "org": "Anthropic",
     "score": 78.7,
     "rank": 5,
     "display": "78.7%"
    }
   ],
   "note": "Best result per model family.",
   "column": "Best score (across scorers)",
   "newest_entry": "2026-06-25",
   "source_id": "epoch",
   "source_name": "Epoch AI Benchmarking Hub",
   "fetched_on": "2026-11-17",
   "stale": false,
   "age_days": 0,
   "benchmark_id": "swe-bench-verified-multimodal",
   "featured": false
  },
  {
   "id": "swebench-verified",
   "title": "SWE-bench Verified (official board)",
   "hint": "SWE-bench Verified",
   "category": "Coding / software engineering",
   "unit": "% resolved",
   "kind": "percent",
   "higher_is_better": true,
   "as_of": "2026-02-26",
   "source_url": "https://www.swebench.com/",
   "data_url": "https://github.com/SWE-bench/swe-bench.github.io",
   "attribution": "SWE-bench team (swebench.com)",
   "rows": [
    {
     "model": "Sonar Foundation Agent + Claude 4.5 Opus",
     "org": "Sonar",
     "score": 79.2,
     "date": "2025-12-05",
     "rank": 1,
     "display": "79.2%"
    },
    {
     "model": "live-SWE-agent + Claude 4.5 Opus medium (20251101)",
     "org": "UIUC",
     "score": 79.2,
     "date": "2025-12-15",
     "rank": 2,
     "display": "79.2%"
    },
    {
     "model": "TRAE + Doubao-Seed-Code",
     "org": "ByteDance",
     "score": 78.8,
     "date": "2025-09-28",
     "rank": 3,
     "display": "78.8%"
    },
    {
     "model": "live-SWE-agent + Gemini 3 Pro Preview (2025-11-18)",
     "org": "UIUC",
     "score": 77.4,
     "date": "2025-11-20",
     "rank": 4,
     "display": "77.4%"
    },
    {
     "model": "EPAM AI/Run Developer Agent v20250719 + Claude 4 Sonnet",
     "org": "EPAM Systems, Inc.",
     "score": 76.8,
     "date": "2025-08-04",
     "rank": 5,
     "display": "76.8%"
    }
   ],
   "note": "Agent + model submissions; date = newest submission on the board.",
   "source_id": "swebench",
   "source_name": "SWE-bench leaderboards",
   "fetched_on": "2026-11-17",
   "stale": false,
   "age_days": 0,
   "benchmark_id": "swe-bench-verified-multimodal",
   "featured": false
  },
  {
   "id": "epoch-cybench",
   "title": "Cybench (unguided)",
   "hint": "Cybench",
   "category": "Cybersecurity",
   "unit": "%",
   "kind": "percent",
   "higher_is_better": true,
   "as_of": "2026-11-17",
   "source_url": "https://epoch.ai/benchmarks",
   "data_url": "https://epoch.ai/data/benchmark_data.zip",
   "attribution": "Epoch AI, Benchmarking Hub (CC BY 4.0)",
   "rows": [
    {
     "model": "Claude Opus 4.6",
     "org": "Anthropic",
     "score": 93.0,
     "rank": 1,
     "display": "93.0%"
    },
    {
     "model": "Claude Opus 4.5",
     "org": "Anthropic",
     "score": 82.0,
     "rank": 2,
     "display": "82.0%"
    },
    {
     "model": "Claude Sonnet 4.5",
     "org": "Anthropic",
     "score": 60.0,
     "rank": 3,
     "display": "60.0%"
    },
    {
     "model": "Grok 4",
     "org": "xAI",
     "score": 43.0,
     "rank": 4,
     "display": "43.0%"
    },
    {
     "model": "Claude Opus 4.1",
     "org": "Anthropic",
     "score": 42.0,
     "rank": 5,
     "display": "42.0%"
    }
   ],
   "note": "Best result per model family.",
   "column": "Unguided % Solved",
   "newest_entry": "2026-02-05",
   "source_id": "epoch",
   "source_name": "Epoch AI Benchmarking Hub",
   "fetched_on": "2026-11-17",
   "stale": false,
   "age_days": 0,
   "benchmark_id": "cybench",
   "featured": false
  },
  {
   "id": "arena-vision",
   "title": "Arena Vision",
   "hint": "Arena (formerly LMArena)",
   "category": "Design, UI & multimodal",
   "unit": "Arena score",
   "kind": "elo",
   "higher_is_better": true,
   "as_of": "2026-11-09",
   "source_url": "https://lmarena.ai/leaderboard/vision",
   "data_url": "https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset",
   "attribution": "Arena / LMArena leaderboard-dataset (CC BY 4.0)",
   "rows": [
    {
     "model": "Claude Fable 5 (High)",
     "org": "anthropic",
     "score": 1325.6,
     "votes": 15536,
     "rank": 1,
     "display": "1326"
    },
    {
     "model": "Claude Opus 5.5 (High)",
     "org": "anthropic",
     "score": 1323.8,
     "votes": 4789,
     "rank": 2,
     "display": "1324"
    },
    {
     "model": "Claude Fable 5.1 (Max)",
     "org": "anthropic",
     "score": 1321.6,
     "votes": 6296,
     "rank": 3,
     "display": "1322"
    },
    {
     "model": "Claude Opus 4.7",
     "org": "anthropic",
     "score": 1318.7,
     "votes": 23777,
     "rank": 4,
     "display": "1319"
    },
    {
     "model": "Claude Opus 5 (High)",
     "org": "anthropic",
     "score": 1317.5,
     "votes": 19345,
     "rank": 5,
     "display": "1318"
    }
   ],
   "note": "Crowd-voted human preference rating (Elo-style).",
   "source_id": "lmarena",
   "source_name": "Arena (LMArena) leaderboard dataset",
   "fetched_on": "2026-11-17",
   "stale": false,
   "age_days": 0,
   "benchmark_id": "arena-formerly-lmarena-text-webdev-vision",
   "featured": false
  },
  {
   "id": "epoch-simpleqa-verified",
   "title": "SimpleQA Verified",
   "hint": "SimpleQA Verified",
   "category": "Knowledge, factuality & hallucination",
   "unit": "%",
   "kind": "percent",
   "higher_is_better": true,
   "as_of": "2026-11-17",
   "source_url": "https://epoch.ai/benchmarks",
   "data_url": "https://epoch.ai/data/benchmark_data.zip",
   "attribution": "Epoch AI, Benchmarking Hub (CC BY 4.0)",
   "rows": [
    {
     "model": "Claude Fable 5.2 (max)",
     "org": "Anthropic",
     "score": 94.0,
     "rank": 1,
     "display": "94.0%"
    },
    {
     "model": "Gemini 4 Carbon (high)",
     "org": "Google DeepMind",
     "score": 77.7,
     "rank": 2,
     "display": "77.7%"
    },
    {
     "model": "Gemini 4 Argon (high)",
     "org": "Google DeepMind",
     "score": 76.0,
     "rank": 3,
     "display": "76.0%"
    },
    {
     "model": "GPT-6 Astra (max)",
     "org": "OpenAI",
     "score": 75.6,
     "rank": 4,
     "display": "75.6%"
    },
    {
     "model": "GPT-6.1 Sol (max)",
     "org": "OpenAI",
     "score": 73.9,
     "rank": 5,
     "display": "73.9%"
    }
   ],
   "note": "Best result per model family.",
   "column": "Best score (across scorers)",
   "newest_entry": "2026-11-06",
   "source_id": "epoch",
   "source_name": "Epoch AI Benchmarking Hub",
   "fetched_on": "2026-11-17",
   "stale": false,
   "age_days": 0,
   "benchmark_id": "simpleqa-verified-facts-suite",
   "featured": false
  },
  {
   "id": "epoch-apex-agents",
   "title": "APEX-Agents",
   "hint": "APEX-Agents",
   "category": "Economic & real-world work value",
   "unit": "%",
   "kind": "percent",
   "higher_is_better": true,
   "as_of": "2026-11-17",
   "source_url": "https://epoch.ai/benchmarks",
   "data_url": "https://epoch.ai/data/benchmark_data.zip",
   "attribution": "Epoch AI, Benchmarking Hub (CC BY 4.0)",
   "rows": [
    {
     "model": "Gemini 4 Carbon",
     "org": "Google DeepMind",
     "score": 84.2,
     "rank": 1,
     "display": "84.2%"
    },
    {
     "model": "Gemini 4 Argon",
     "org": "Google DeepMind",
     "score": 82.4,
     "rank": 2,
     "display": "82.4%"
    },
    {
     "model": "Claude Fable 5.2",
     "org": "Anthropic",
     "score": 80.1,
     "rank": 3,
     "display": "80.1%"
    },
    {
     "model": "Claude Sonnet 5.5 (max)",
     "org": "Anthropic",
     "score": 75.5,
     "rank": 4,
     "display": "75.5%"
    },
    {
     "model": "Claude Opus 5.5 (max)",
     "org": "Anthropic",
     "score": 73.5,
     "rank": 5,
     "display": "73.5%"
    }
   ],
   "note": "Best result per model family.",
   "column": "Pass@1 score",
   "newest_entry": "2026-11-10",
   "source_id": "epoch",
   "source_name": "Epoch AI Benchmarking Hub",
   "fetched_on": "2026-11-17",
   "stale": false,
   "age_days": 0,
   "benchmark_id": "apex-agents-mercor",
   "featured": false
  },
  {
   "id": "epoch-gdpval",
   "title": "GDPval (win rate)",
   "hint": "GDPval",
   "category": "Economic & real-world work value",
   "unit": "%",
   "kind": "percent",
   "higher_is_better": true,
   "as_of": "2026-11-17",
   "source_url": "https://epoch.ai/benchmarks",
   "data_url": "https://epoch.ai/data/benchmark_data.zip",
   "attribution": "Epoch AI, Benchmarking Hub (CC BY 4.0)",
   "rows": [
    {
     "model": "GPT-5.2 (none)",
     "org": "OpenAI",
     "score": 49.7,
     "rank": 1,
     "display": "49.7%"
    },
    {
     "model": "Claude Opus 4.5 (no thinking)",
     "org": "Anthropic",
     "score": 45.5,
     "rank": 2,
     "display": "45.5%"
    },
    {
     "model": "Claude Opus 4.1",
     "org": "Anthropic",
     "score": 43.6,
     "rank": 3,
     "display": "43.6%"
    },
    {
     "model": "Claude Sonnet 4.5 (no thinking)",
     "org": "Anthropic",
     "score": 42.5,
     "rank": 4,
     "display": "42.5%"
    },
    {
     "model": "Gemini 3 Pro Preview",
     "org": "Google DeepMind",
     "score": 40.3,
     "rank": 5,
     "display": "40.3%"
    }
   ],
   "note": "Best result per model family.",
   "column": "Win Rate (%)",
   "newest_entry": "2025-12-11",
   "source_id": "epoch",
   "source_name": "Epoch AI Benchmarking Hub",
   "fetched_on": "2026-11-17",
   "stale": false,
   "age_days": 0,
   "benchmark_id": "gdpval-gdpval-aa-v2-1",
   "featured": false
  },
  {
   "id": "epoch-rli",
   "title": "Remote Labor Index",
   "hint": "Remote Labor Index",
   "category": "Economic & real-world work value",
   "unit": "%",
   "kind": "percent",
   "higher_is_better": true,
   "as_of": "2026-11-17",
   "source_url": "https://epoch.ai/benchmarks",
   "data_url": "https://epoch.ai/data/benchmark_data.zip",
   "attribution": "Epoch AI, Benchmarking Hub (CC BY 4.0)",
   "rows": [
    {
     "model": "Claude Fable 5.2",
     "org": "Anthropic",
     "score": 41.3,
     "rank": 1,
     "display": "41.3%"
    },
    {
     "model": "Claude Opus 5.5",
     "org": "Anthropic",
     "score": 21.2,
     "rank": 2,
     "display": "21.2%"
    },
    {
     "model": "GPT-6 Astra",
     "org": "OpenAI",
     "score": 20.8,
     "rank": 3,
     "display": "20.8%"
    },
    {
     "model": "Claude Fable 5.1",
     "org": "Anthropic",
     "score": 17.9,
     "rank": 4,
     "display": "17.9%"
    },
    {
     "model": "Gemini 4 Argon",
     "org": "Google DeepMind",
     "score": 17.1,
     "rank": 5,
     "display": "17.1%"
    }
   ],
   "note": "Best result per model family.",
   "column": "Score",
   "newest_entry": "2026-11-09",
   "source_id": "epoch",
   "source_name": "Epoch AI Benchmarking Hub",
   "fetched_on": "2026-11-17",
   "stale": false,
   "age_days": 0,
   "benchmark_id": "remote-labor-index",
   "featured": false
  }
 ],
 "benchmarks": [
  {
   "id": "artificial-analysis-intelligence-index-v4-3-2",
   "name": "Artificial Analysis Intelligence Index v4.3.2",
   "tier": 1,
   "status": "active",
   "status_note": "Active",
   "what": "Combined score from 10 evals (agents 30%, coding 20%, sci. reasoning 20%, general 30%)",
   "urls": [
    {
     "url": "https://artificialanalysis.ai/methodology/intelligence-benchmarking",
     "label": "artificialanalysis.ai/methodology/intelligence-benchmarking",
     "third_party": false
    }
   ],
   "categories": [
    "Aggregators"
   ],
   "category": "Aggregators"
  },
  {
   "id": "epoch-capabilities-index-eci-benchmarking-hub",
   "name": "Epoch Capabilities Index (ECI) + Benchmarking Hub",
   "tier": 1,
   "status": "active",
   "status_note": "Active",
   "what": "Capability scale built with item response theory (IRT) from about 61 benchmarks; hub for independent runs",
   "urls": [
    {
     "url": "https://epoch.ai/eci",
     "label": "epoch.ai/eci",
     "third_party": false
    },
    {
     "url": "https://epoch.ai/benchmarks",
     "label": "epoch.ai/benchmarks",
     "third_party": false
    }
   ],
   "categories": [
    "Aggregators"
   ],
   "category": "Aggregators"
  },
  {
   "id": "humanity-s-last-exam-hle-hle-rolling",
   "name": "Humanity's Last Exam (HLE / HLE-Rolling)",
   "tier": 1,
   "status": "active",
   "status_note": "Active (HLE-Rolling released Sep 2026 as the migration path)",
   "what": "2,500 expert questions at the edge of academic knowledge",
   "urls": [
    {
     "url": "https://lastexam.ai/",
     "label": "lastexam.ai",
     "third_party": false
    },
    {
     "url": "https://labs.scale.com/leaderboard/humanitys_last_exam",
     "label": "labs.scale.com/leaderboard/humanitys_last_exam",
     "third_party": false
    }
   ],
   "categories": [
    "General intelligence & reasoning"
   ],
   "category": "General intelligence & reasoning"
  },
  {
   "id": "arc-agi-3-with-arc-agi-2",
   "name": "ARC-AGI-3 (with ARC-AGI-2)",
   "tier": 1,
   "status": "active",
   "status_note": "Active (v3 launched 25 Mar 2026)",
   "what": "Interactive novel-environment learning, fluid intelligence",
   "urls": [
    {
     "url": "https://arcprize.org/leaderboard",
     "label": "arcprize.org/leaderboard",
     "third_party": false
    },
    {
     "url": "https://arcprize.org/blog/arc-agi-3-launch",
     "label": "arcprize.org/blog/arc-agi-3-launch",
     "third_party": false
    }
   ],
   "categories": [
    "General intelligence & reasoning"
   ],
   "category": "General intelligence & reasoning"
  },
  {
   "id": "swe-bench-pro-scale",
   "name": "SWE-bench Pro (Scale)",
   "tier": 1,
   "status": "active",
   "status_note": "Active (successor to Verified)",
   "what": "Long-horizon, multi-file, multi-language real repo issues",
   "urls": [
    {
     "url": "https://labs.scale.com/leaderboard/swe_bench_pro_public",
     "label": "labs.scale.com/leaderboard/swe_bench_pro_public",
     "third_party": false
    }
   ],
   "categories": [
    "Coding / software engineering"
   ],
   "category": "Coding / software engineering"
  },
  {
   "id": "terminal-bench-4-0-and-3-0",
   "name": "Terminal-Bench 4.0 (and 3.0)",
   "tier": 1,
   "status": "active",
   "status_note": "Active (4.0 on 28 Aug 2026; 2.0/2.1 superseded)",
   "what": "Agentic terminal tasks: software engineering, sysadmin, ML, security",
   "urls": [
    {
     "url": "https://www.tbench.ai/news",
     "label": "tbench.ai/news",
     "third_party": false
    },
    {
     "url": "https://www.tbench.ai/leaderboard/terminal-bench/2.0",
     "label": "tbench.ai/leaderboard/terminal-bench/2.0",
     "third_party": false
    }
   ],
   "categories": [
    "Coding / software engineering"
   ],
   "category": "Coding / software engineering"
  },
  {
   "id": "gdpval-gdpval-aa-v2-1",
   "name": "GDPval / GDPval-AA v2.1",
   "tier": 1,
   "status": "active",
   "status_note": "Active via AA (OpenAI-hosted board inactive)",
   "what": "Expert-judged deliverables across 44 occupations",
   "urls": [
    {
     "url": "https://openai.com/index/gdpval/",
     "label": "openai.com/index/gdpval",
     "third_party": false
    },
    {
     "url": "https://artificialanalysis.ai/evaluations/gdpval-aa?models=",
     "label": "artificialanalysis.ai/evaluations/gdpval-aa?models=",
     "third_party": false
    }
   ],
   "categories": [
    "Economic & real-world work value"
   ],
   "category": "Economic & real-world work value"
  },
  {
   "id": "frontiermath-tiers-1-4-v2",
   "name": "FrontierMath Tiers 1–4 (v2)",
   "tier": 1,
   "status": "near-saturated",
   "status_note": "Near-sat. at Tier 4 (reported); Open Problems is the next step",
   "what": "Research-level math, private problems",
   "urls": [
    {
     "url": "https://epoch.ai/frontiermath/tiers-1-4",
     "label": "epoch.ai/frontiermath/tiers-1-4",
     "third_party": false
    },
    {
     "url": "https://epoch.ai/benchmarks/frontiermath-tier-4-v2",
     "label": "epoch.ai/benchmarks/frontiermath-tier-4-v2",
     "third_party": false
    }
   ],
   "categories": [
    "Math"
   ],
   "category": "Math"
  },
  {
   "id": "osworld-2-0-and-osworld-verified",
   "name": "OSWorld 2.0 (and OSWorld-Verified)",
   "tier": 1,
   "status": "active",
   "status_note": "Active (2.0/2.1 supersede Verified)",
   "what": "Long-horizon real computer-use workflows (about 1.6 h median per task)",
   "urls": [
    {
     "url": "https://osworld-v2.xlang.ai/",
     "label": "osworld-v2.xlang.ai",
     "third_party": false
    },
    {
     "url": "https://github.com/xlang-ai/OSWorld-V2",
     "label": "github.com/xlang-ai/OSWorld-V2",
     "third_party": false
    }
   ],
   "categories": [
    "Agentic, computer use & long-horizon"
   ],
   "category": "Agentic, computer use & long-horizon"
  },
  {
   "id": "arena-formerly-lmarena-text-webdev-vision",
   "name": "Arena (formerly LMArena) Text/WebDev/Vision",
   "tier": 1,
   "status": "active",
   "status_note": "Active (renamed to Arena on 28 Jan 2026)",
   "what": "Crowd-voted human-preference Elo",
   "urls": [
    {
     "url": "https://lmarena.ai/leaderboard",
     "label": "lmarena.ai/leaderboard",
     "third_party": false
    },
    {
     "url": "https://arena.ai/blog/lmarena-is-now-arena/",
     "label": "arena.ai/blog/lmarena-is-now-arena",
     "third_party": false
    }
   ],
   "categories": [
    "Aggregators",
    "Design, UI & multimodal"
   ],
   "category": "Aggregators"
  },
  {
   "id": "tau3-bench-tau3-was-tau2",
   "name": "τ³-bench (tau3; was τ²)",
   "tier": 2,
   "status": "active",
   "status_note": "Active (supersedes τ²)",
   "what": "Tool-agent-user dialogue, incl. banking knowledge and voice",
   "urls": [
    {
     "url": "https://taubench.com/leaderboard",
     "label": "taubench.com/leaderboard",
     "third_party": false
    }
   ],
   "categories": [
    "Agentic, computer use & long-horizon"
   ],
   "category": "Agentic, computer use & long-horizon"
  },
  {
   "id": "remote-labor-index",
   "name": "Remote Labor Index",
   "tier": 2,
   "status": "active",
   "status_note": "Active",
   "what": "Share of real freelance projects completed to client standard",
   "urls": [
    {
     "url": "https://labs.scale.com/leaderboard/rli",
     "label": "labs.scale.com/leaderboard/rli",
     "third_party": false
    },
    {
     "url": "https://www.remotelabor.ai/",
     "label": "remotelabor.ai",
     "third_party": false
    }
   ],
   "categories": [
    "Economic & real-world work value"
   ],
   "category": "Economic & real-world work value"
  },
  {
   "id": "vending-bench-2",
   "name": "Vending-Bench 2",
   "tier": 2,
   "status": "active",
   "status_note": "Active",
   "what": "Year-long autonomous business operation (final bank balance)",
   "urls": [
    {
     "url": "https://andonlabs.com/evals/vending-bench-2",
     "label": "andonlabs.com/evals/vending-bench-2",
     "third_party": false
    }
   ],
   "categories": [
    "Agentic, computer use & long-horizon",
    "Economic & real-world work value"
   ],
   "category": "Agentic, computer use & long-horizon"
  },
  {
   "id": "metr-time-horizon-1-1",
   "name": "METR Time Horizon 1.1",
   "tier": 2,
   "status": "frozen",
   "status_note": "Frozen (page not updated since 8 Sep 2026)",
   "what": "Length of task (in human time) an agent completes at 50%",
   "urls": [
    {
     "url": "https://metr.org/time-horizons/",
     "label": "metr.org/time-horizons",
     "third_party": false
    }
   ],
   "categories": [
    "Agentic, computer use & long-horizon"
   ],
   "category": "Agentic, computer use & long-horizon"
  },
  {
   "id": "uk-aisi-cyber-ranges",
   "name": "UK AISI cyber ranges",
   "tier": 2,
   "status": "active",
   "status_note": "Active (reports, not a public board)",
   "what": "Multi-step network/ICS attack simulations (32-step and 7-step)",
   "urls": [
    {
     "url": "https://www.aisi.gov.uk/blog/how-do-frontier-ai-agents-perform-in-multi-step-cyber-attack-scenarios",
     "label": "aisi.gov.uk/blog/how-do-frontier-ai-agents-perform-in-mul…",
     "third_party": false
    }
   ],
   "categories": [
    "Cybersecurity"
   ],
   "category": "Cybersecurity"
  },
  {
   "id": "cybergym",
   "name": "CyberGym",
   "tier": 2,
   "status": "near-saturated",
   "status_note": "Near-sat. (third-party reports)",
   "what": "Reproducing real vulnerabilities with working proofs of concept (PoCs)",
   "urls": [
    {
     "url": "https://www.cybergym.io/cybergym/",
     "label": "cybergym.io/cybergym",
     "third_party": false
    }
   ],
   "categories": [
    "Cybersecurity"
   ],
   "category": "Cybersecurity"
  },
  {
   "id": "cybench",
   "name": "Cybench",
   "tier": 2,
   "status": "near-saturated",
   "status_note": "Near-sat.",
   "what": "40 professional CTF tasks",
   "urls": [
    {
     "url": "https://cybench.github.io/index.html",
     "label": "cybench.github.io/index.html",
     "third_party": false
    }
   ],
   "categories": [
    "Cybersecurity"
   ],
   "category": "Cybersecurity"
  },
  {
   "id": "livecodebench-pro",
   "name": "LiveCodeBench Pro",
   "tier": 2,
   "status": "active",
   "status_note": "Active",
   "what": "Contest programming (Elo), post-cutoff problems",
   "urls": [
    {
     "url": "https://atlas.kevinhu.io/benchmarks/livecodebench-pro",
     "label": "atlas.kevinhu.io/benchmarks/livecodebench-pro",
     "third_party": true
    }
   ],
   "categories": [
    "Coding / software engineering"
   ],
   "category": "Coding / software engineering"
  },
  {
   "id": "browsecomp-plus",
   "name": "BrowseComp-Plus",
   "tier": 2,
   "status": "active",
   "status_note": "Active (original BrowseComp near-sat.)",
   "what": "Deep-research agent on a fixed verified corpus",
   "urls": [
    {
     "url": "https://texttron.github.io/BrowseComp-Plus/",
     "label": "texttron.github.io/BrowseComp-Plus",
     "third_party": false
    }
   ],
   "categories": [
    "Agentic, computer use & long-horizon"
   ],
   "category": "Agentic, computer use & long-horizon"
  },
  {
   "id": "mmmu-pro",
   "name": "MMMU-Pro",
   "tier": 2,
   "status": "near-saturated",
   "status_note": "Near-sat.",
   "what": "Expert multimodal (image) reasoning",
   "urls": [
    {
     "url": "https://benchmarklist.com/benchmarks/mmmu_pro/",
     "label": "benchmarklist.com/benchmarks/mmmu_pro",
     "third_party": true
    }
   ],
   "categories": [
    "Design, UI & multimodal"
   ],
   "category": "Design, UI & multimodal"
  },
  {
   "id": "screenspot-pro",
   "name": "ScreenSpot-Pro",
   "tier": 2,
   "status": "active",
   "status_note": "Active",
   "what": "GUI grounding on professional high-res screens",
   "urls": [
    {
     "url": "https://gui-agent.github.io/grounding-leaderboard/index.html",
     "label": "gui-agent.github.io/grounding-leaderboard/index.html",
     "third_party": false
    }
   ],
   "categories": [
    "Agentic, computer use & long-horizon",
    "Design, UI & multimodal"
   ],
   "category": "Agentic, computer use & long-horizon"
  },
  {
   "id": "design-arena",
   "name": "Design Arena",
   "tier": 2,
   "status": "active",
   "status_note": "Active",
   "what": "Crowd-voted web, UI, 3D and graphic design",
   "urls": [
    {
     "url": "https://www.designarena.ai/leaderboard",
     "label": "designarena.ai/leaderboard",
     "third_party": false
    }
   ],
   "categories": [
    "Design, UI & multimodal"
   ],
   "category": "Design, UI & multimodal"
  },
  {
   "id": "arena-webdev",
   "name": "Arena WebDev",
   "tier": 2,
   "status": "active",
   "status_note": "Active",
   "what": "Human preference for generated web apps",
   "urls": [
    {
     "url": "https://lmarena.ai/leaderboard/code",
     "label": "lmarena.ai/leaderboard/code",
     "third_party": false
    }
   ],
   "categories": [
    "Coding / software engineering",
    "Design, UI & multimodal"
   ],
   "category": "Coding / software engineering"
  },
  {
   "id": "aa-omniscience",
   "name": "AA-Omniscience",
   "tier": 2,
   "status": "active",
   "status_note": "Active",
   "what": "Knowledge plus hallucination penalty (abstaining beats guessing)",
   "urls": [
    {
     "url": "https://artificialanalysis.ai/methodology/intelligence-benchmarking",
     "label": "artificialanalysis.ai/methodology/intelligence-benchmarking",
     "third_party": false
    }
   ],
   "categories": [
    "Knowledge, factuality & hallucination"
   ],
   "category": "Knowledge, factuality & hallucination"
  },
  {
   "id": "simpleqa-verified-facts-suite",
   "name": "SimpleQA Verified / FACTS suite",
   "tier": 2,
   "status": "active",
   "status_note": "Active",
   "what": "Short-form factuality; grounding, parametric, search, multimodal",
   "urls": [
    {
     "url": "https://www.kaggle.com/benchmarks/deepmind/simpleqa-verified",
     "label": "kaggle.com/benchmarks/deepmind/simpleqa-verified",
     "third_party": false
    },
    {
     "url": "https://www.kaggle.com/benchmarks/google/facts",
     "label": "kaggle.com/benchmarks/google/facts",
     "third_party": false
    }
   ],
   "categories": [
    "Knowledge, factuality & hallucination"
   ],
   "category": "Knowledge, factuality & hallucination"
  },
  {
   "id": "ifbench",
   "name": "IFBench",
   "tier": 2,
   "status": "active",
   "status_note": "Active",
   "what": "Generalising to 58 unseen verifiable constraints",
   "urls": [
    {
     "url": "https://artificialanalysis.ai/evaluations/ifbench",
     "label": "artificialanalysis.ai/evaluations/ifbench",
     "third_party": false
    }
   ],
   "categories": [
    "Instruction following"
   ],
   "category": "Instruction following"
  },
  {
   "id": "mrcr-v2-aa-lcr-v1-1",
   "name": "MRCR v2 / AA-LCR v1.1",
   "tier": 2,
   "status": "active",
   "status_note": "Active (MRCR 8-needle near-sat. at top)",
   "what": "Long-context retrieval and reasoning (about 100k+ tokens)",
   "urls": [
    {
     "url": "https://huggingface.co/datasets/openai/mrcr",
     "label": "huggingface.co/datasets/openai/mrcr",
     "third_party": false
    },
    {
     "url": "https://huggingface.co/datasets/ArtificialAnalysis/AA-LCR",
     "label": "huggingface.co/datasets/ArtificialAnalysis/AA-LCR",
     "third_party": false
    }
   ],
   "categories": [
    "Long context"
   ],
   "category": "Long context"
  },
  {
   "id": "imo-proofbench-matharena-proofs",
   "name": "IMO-ProofBench / MathArena proofs",
   "tier": 2,
   "status": "active",
   "status_note": "Active",
   "what": "Olympiad proof writing, graded 0–7",
   "urls": [
    {
     "url": "https://github.com/EnvCommons/IMO-Bench",
     "label": "github.com/EnvCommons/IMO-Bench",
     "third_party": false
    },
    {
     "url": "https://matharena.ai/no_final_answer/",
     "label": "matharena.ai/no_final_answer",
     "third_party": false
    }
   ],
   "categories": [
    "Math"
   ],
   "category": "Math"
  },
  {
   "id": "critpt",
   "name": "CritPt",
   "tier": 2,
   "status": "active",
   "status_note": "Active",
   "what": "Unpublished research-level physics",
   "urls": [
    {
     "url": "https://critpt.com/",
     "label": "critpt.com",
     "third_party": false
    }
   ],
   "categories": [
    "General intelligence & reasoning"
   ],
   "category": "General intelligence & reasoning"
  },
  {
   "id": "apex-agents-mercor",
   "name": "APEX-Agents (Mercor)",
   "tier": 2,
   "status": "active",
   "status_note": "Active",
   "what": "Banking, consulting and legal cross-app agent work",
   "urls": [
    {
     "url": "https://arxiv.org/abs/2601.14242",
     "label": "arxiv.org/abs/2601.14242",
     "third_party": false
    }
   ],
   "categories": [
    "Economic & real-world work value"
   ],
   "category": "Economic & real-world work value"
  },
  {
   "id": "livebench",
   "name": "LiveBench",
   "tier": 2,
   "status": "active",
   "status_note": "Active",
   "what": "Monthly-refreshed, contamination-limited general suite",
   "urls": [
    {
     "url": "https://livebench.ai/",
     "label": "livebench.ai",
     "third_party": false
    }
   ],
   "categories": [
    "Aggregators"
   ],
   "category": "Aggregators"
  },
  {
   "id": "scale-seal-leaderboards",
   "name": "Scale SEAL leaderboards",
   "tier": 2,
   "status": "active",
   "status_note": "Active",
   "what": "Private expert-graded evals (several categories)",
   "urls": [
    {
     "url": "https://scale.com/blog/leaderboard",
     "label": "scale.com/blog/leaderboard",
     "third_party": false
    }
   ],
   "categories": [
    "Aggregators"
   ],
   "category": "Aggregators"
  },
  {
   "id": "euroeval-formerly-scandeval-swedish-scandinavian",
   "name": "EuroEval (formerly ScandEval): Swedish/Scandinavian",
   "tier": 2,
   "status": "active",
   "status_note": "Active",
   "what": "Nordic-language NLU/NLG/knowledge",
   "urls": [
    {
     "url": "https://euroeval.com/leaderboards/swedish",
     "label": "euroeval.com/leaderboards/swedish",
     "third_party": false
    },
    {
     "url": "https://euroeval.com/leaderboards/scandinavian",
     "label": "euroeval.com/leaderboards/scandinavian",
     "third_party": false
    }
   ],
   "categories": [
    "Multilingual (Swedish/Nordic)"
   ],
   "category": "Multilingual (Swedish/Nordic)"
  },
  {
   "id": "gpqa-diamond",
   "name": "GPQA Diamond",
   "tier": 3,
   "status": "saturated",
   "status_note": "Saturated",
   "what": "Graduate \"Google-proof\" science MCQ",
   "urls": [
    {
     "url": "https://arxiv.org/abs/2311.12022",
     "label": "arxiv.org/abs/2311.12022",
     "third_party": false
    }
   ],
   "categories": [
    "General intelligence & reasoning"
   ],
   "category": "General intelligence & reasoning"
  },
  {
   "id": "mmlu-pro",
   "name": "MMLU-Pro",
   "tier": 3,
   "status": "near-saturated",
   "status_note": "Near-sat.",
   "what": "Harder multi-task knowledge MCQ",
   "urls": [
    {
     "url": "https://arxiv.org/html/2406.01574v5",
     "label": "arxiv.org/html/2406.01574v5",
     "third_party": false
    }
   ],
   "categories": [
    "General intelligence & reasoning"
   ],
   "category": "General intelligence & reasoning"
  },
  {
   "id": "swe-bench-verified-multimodal",
   "name": "SWE-bench Verified / Multimodal",
   "tier": 3,
   "status": "saturated",
   "status_note": "Saturated (Verified); superseded by Pro",
   "what": "500 Python GitHub issues / visual issues",
   "urls": [
    {
     "url": "https://www.swebench.com/",
     "label": "swebench.com",
     "third_party": false
    }
   ],
   "categories": [
    "Coding / software engineering"
   ],
   "category": "Coding / software engineering"
  },
  {
   "id": "aime-hmmt-matharena",
   "name": "AIME / HMMT (MathArena)",
   "tier": 3,
   "status": "saturated",
   "status_note": "Saturated (MathArena dropped final-answer sets)",
   "what": "Final-answer competition math",
   "urls": [
    {
     "url": "https://matharena.ai/competitions",
     "label": "matharena.ai/competitions",
     "third_party": false
    }
   ],
   "categories": [
    "Math"
   ],
   "category": "Math"
  },
  {
   "id": "frontiermath-open-problems",
   "name": "FrontierMath: Open Problems",
   "tier": 3,
   "status": "active",
   "status_note": "Active (no score ranking)",
   "what": "Unsolved research math (records first solutions)",
   "urls": [
    {
     "url": "https://epoch.ai/frontiermath/open-problems",
     "label": "epoch.ai/frontiermath/open-problems",
     "third_party": false
    }
   ],
   "categories": [
    "Math"
   ],
   "category": "Math"
  },
  {
   "id": "aider-polyglot",
   "name": "Aider Polyglot",
   "tier": 3,
   "status": "stale",
   "status_note": "Stale or near-sat. (board looks unmaintained)",
   "what": "225 Exercism edits in 6 languages",
   "urls": [
    {
     "url": "https://aider.chat/docs/leaderboards/",
     "label": "aider.chat/docs/leaderboards",
     "third_party": false
    }
   ],
   "categories": [
    "Coding / software engineering"
   ],
   "category": "Coding / software engineering"
  },
  {
   "id": "gaia",
   "name": "GAIA",
   "tier": 3,
   "status": "saturated",
   "status_note": "Saturated",
   "what": "General assistant tool-use questions",
   "urls": [
    {
     "url": "https://personalagentbench.com/benchmarks/gaia/",
     "label": "personalagentbench.com/benchmarks/gaia",
     "third_party": true
    }
   ],
   "categories": [
    "Agentic, computer use & long-horizon"
   ],
   "category": "Agentic, computer use & long-horizon"
  },
  {
   "id": "webarena-verified",
   "name": "WebArena-Verified",
   "tier": 3,
   "status": "active",
   "status_note": "Active",
   "what": "Reproducible web-agent tasks",
   "urls": [
    {
     "url": "https://github.com/ServiceNow/webarena-verified",
     "label": "github.com/ServiceNow/webarena-verified",
     "third_party": false
    }
   ],
   "categories": [
    "Agentic, computer use & long-horizon"
   ],
   "category": "Agentic, computer use & long-horizon"
  },
  {
   "id": "browsecomp-original",
   "name": "BrowseComp (original)",
   "tier": 3,
   "status": "near-saturated",
   "status_note": "Near-sat.",
   "what": "Hard-to-find web facts",
   "urls": [
    {
     "url": "https://leaderboard.steel.dev/leaderboards/browsecomp/",
     "label": "leaderboard.steel.dev/leaderboards/browsecomp",
     "third_party": true
    }
   ],
   "categories": [
    "Agentic, computer use & long-horizon"
   ],
   "category": "Agentic, computer use & long-horizon",
   "category_heuristic": true
  },
  {
   "id": "mle-bench",
   "name": "MLE-bench",
   "tier": 3,
   "status": "frozen",
   "status_note": "Frozen",
   "what": "Kaggle-style ML engineering",
   "urls": [
    {
     "url": "https://github.com/openai/mle-bench/",
     "label": "github.com/openai/mle-bench",
     "third_party": false
    }
   ],
   "categories": [
    "Agentic, computer use & long-horizon"
   ],
   "category": "Agentic, computer use & long-horizon"
  },
  {
   "id": "swe-lancer",
   "name": "SWE-Lancer",
   "tier": 3,
   "status": "frozen",
   "status_note": "Frozen or stale",
   "what": "Paid freelance software tasks",
   "urls": [
    {
     "url": "https://swelancer.github.io/leaderboard/",
     "label": "swelancer.github.io/leaderboard",
     "third_party": false
    }
   ],
   "categories": [
    "Coding / software engineering",
    "Economic & real-world work value"
   ],
   "category": "Coding / software engineering"
  },
  {
   "id": "paperbench",
   "name": "PaperBench",
   "tier": 3,
   "status": "near-saturated",
   "status_note": "Near-sat. (reported)",
   "what": "Replicating ML research papers",
   "urls": [
    {
     "url": "https://benchlm.ai/benchmarks/paperbench",
     "label": "benchlm.ai/benchmarks/paperbench",
     "third_party": true
    }
   ],
   "categories": [
    "Agentic, computer use & long-horizon"
   ],
   "category": "Agentic, computer use & long-horizon",
   "category_heuristic": true
  },
  {
   "id": "scicode",
   "name": "SciCode",
   "tier": 3,
   "status": "active",
   "status_note": "Active (AA index component)",
   "what": "Scientific-computing code",
   "urls": [
    {
     "url": "https://scicode-bench.github.io/",
     "label": "scicode-bench.github.io",
     "third_party": false
    }
   ],
   "categories": [
    "Coding / software engineering"
   ],
   "category": "Coding / software engineering"
  },
  {
   "id": "automationbench-zapier",
   "name": "AutomationBench (Zapier)",
   "tier": 3,
   "status": "active",
   "status_note": "Active",
   "what": "Multi-app SaaS workflow automation",
   "urls": [
    {
     "url": "https://zapier.com/benchmarks",
     "label": "zapier.com/benchmarks",
     "third_party": false
    }
   ],
   "categories": [
    "Agentic, computer use & long-horizon"
   ],
   "category": "Agentic, computer use & long-horizon"
  },
  {
   "id": "terminal-bench-science-0-1-harbor-index",
   "name": "Terminal-Bench-Science 0.1 / Harbor-Index",
   "tier": 3,
   "status": "new",
   "status_note": "New",
   "what": "Scientific workflows / compact agent index",
   "urls": [
    {
     "url": "https://www.tbench.ai/news",
     "label": "tbench.ai/news",
     "third_party": false
    }
   ],
   "categories": [
    "Coding / software engineering"
   ],
   "category": "Coding / software engineering",
   "category_heuristic": true
  },
  {
   "id": "cyberseceval-4-cybersoceval-autopatchbench",
   "name": "CyberSecEval 4 (+CyberSOCEval, AutoPatchBench)",
   "tier": 3,
   "status": "active",
   "status_note": "Active",
   "what": "Cyber risk and defensive SOC/patching",
   "urls": [
    {
     "url": "https://meta-llama.github.io/PurpleLlama/CyberSecEval/docs/intro",
     "label": "meta-llama.github.io/PurpleLlama/CyberSecEval/docs/intro",
     "third_party": false
    }
   ],
   "categories": [
    "Cybersecurity"
   ],
   "category": "Cybersecurity"
  },
  {
   "id": "cve-bench",
   "name": "CVE-Bench",
   "tier": 3,
   "status": "active",
   "status_note": "Active (no live board found)",
   "what": "Exploiting critical web-app CVEs",
   "urls": [
    {
     "url": "https://dl.acm.org/doi/abs/10.5555/3780338.3783551",
     "label": "dl.acm.org/doi/abs/10.5555/3780338.3783551",
     "third_party": false
    }
   ],
   "categories": [
    "Cybersecurity"
   ],
   "category": "Cybersecurity"
  },
  {
   "id": "bountybench",
   "name": "BountyBench",
   "tier": 3,
   "status": "active",
   "status_note": "Active",
   "what": "Detect, exploit and patch on real bug bounties (dollar value)",
   "urls": [
    {
     "url": "https://arxiv.org/html/2505.15216v3",
     "label": "arxiv.org/html/2505.15216v3",
     "third_party": false
    }
   ],
   "categories": [
    "Cybersecurity"
   ],
   "category": "Cybersecurity"
  },
  {
   "id": "sec-bench-sec-bench-pro",
   "name": "SEC-bench / SEC-bench Pro",
   "tier": 3,
   "status": "active",
   "status_note": "Active",
   "what": "CVE proof-of-concept generation and patching (\"SWE-bench for security\")",
   "urls": [
    {
     "url": "https://github.com/SEC-bench/SEC-bench-Pro",
     "label": "github.com/SEC-bench/SEC-bench-Pro",
     "third_party": false
    }
   ],
   "categories": [
    "Cybersecurity"
   ],
   "category": "Cybersecurity"
  },
  {
   "id": "ctibench",
   "name": "CTIBench",
   "tier": 3,
   "status": "active",
   "status_note": "Active",
   "what": "Threat-intelligence knowledge, CVE→CWE, CVSS, ATT&CK",
   "urls": [
    {
     "url": "https://github.com/maveryn/cti-bench",
     "label": "github.com/maveryn/cti-bench",
     "third_party": false
    }
   ],
   "categories": [
    "Cybersecurity"
   ],
   "category": "Cybersecurity"
  },
  {
   "id": "wmdp-cyber-subset",
   "name": "WMDP (Cyber subset)",
   "tier": 3,
   "status": "static",
   "status_note": "Static (2024)",
   "what": "Hazardous cyber, bio and chem knowledge proxy; unlearning",
   "urls": [
    {
     "url": "https://www.wmdp.ai/",
     "label": "wmdp.ai",
     "third_party": false
    }
   ],
   "categories": [
    "Cybersecurity",
    "Safety, alignment & robustness"
   ],
   "category": "Cybersecurity"
  },
  {
   "id": "mlcommons-ailuminate",
   "name": "MLCommons AILuminate",
   "tier": 3,
   "status": "active",
   "status_note": "Active (results dated)",
   "what": "Chat safety across 12 hazard categories",
   "urls": [
    {
     "url": "https://mlcommons.org/ailuminate/safety/",
     "label": "mlcommons.org/ailuminate/safety",
     "third_party": false
    }
   ],
   "categories": [
    "Safety, alignment & robustness"
   ],
   "category": "Safety, alignment & robustness"
  },
  {
   "id": "agentharm-helm-safety",
   "name": "AgentHarm / HELM Safety",
   "tier": 3,
   "status": "active",
   "status_note": "Active (third-party boards)",
   "what": "Harmful agent compliance / safety suite",
   "urls": [
    {
     "url": "https://benchmarklist.com/benchmarks/agentharm/",
     "label": "benchmarklist.com/benchmarks/agentharm",
     "third_party": false
    },
    {
     "url": "https://benchmarklist.com/benchmarks/helm_safety/",
     "label": "benchmarklist.com/benchmarks/helm_safety",
     "third_party": false
    }
   ],
   "categories": [
    "Safety, alignment & robustness"
   ],
   "category": "Safety, alignment & robustness"
  },
  {
   "id": "backdoorllm",
   "name": "BackdoorLLM",
   "tier": 3,
   "status": "static",
   "status_note": "Static (NeurIPS 2025, no live board)",
   "what": "Backdoor attacks and defences on LLMs (data, weight, CoT poisoning)",
   "urls": [
    {
     "url": "https://bboylyg.github.io/backdoorllm-website.github.io/",
     "label": "bboylyg.github.io/backdoorllm-website.github.io",
     "third_party": false
    }
   ],
   "categories": [
    "Safety, alignment & robustness"
   ],
   "category": "Safety, alignment & robustness"
  },
  {
   "id": "poisonforge",
   "name": "PoisonForge",
   "tier": 3,
   "status": "new",
   "status_note": "New (third-party board only)",
   "what": "Targeted task-level poisoning of instruction-tuned LLMs",
   "urls": [
    {
     "url": "https://benchmarklist.com/benchmarks/poisonforge_task_level_targeted_poisoning_benchmark_for_instruction_tuned_llms/",
     "label": "benchmarklist.com/benchmarks/poisonforge_task_level_targe…",
     "third_party": false
    }
   ],
   "categories": [
    "Safety, alignment & robustness"
   ],
   "category": "Safety, alignment & robustness"
  },
  {
   "id": "video-mme-video-mme-v2",
   "name": "Video-MME / Video-MME-v2",
   "tier": 3,
   "status": "active",
   "status_note": "Active (v2 supersedes v1)",
   "what": "Video understanding",
   "urls": [
    {
     "url": "https://video-mme.github.io/home%5Fpage.html",
     "label": "video-mme.github.io/home%5Fpage.html",
     "third_party": false
    },
    {
     "url": "https://www.alphaxiv.org/abs/2604.05015",
     "label": "alphaxiv.org/abs/2604.05015",
     "third_party": false
    }
   ],
   "categories": [
    "Design, UI & multimodal"
   ],
   "category": "Design, UI & multimodal"
  },
  {
   "id": "image-arenas-arena-t2i-aa-image-arena",
   "name": "Image arenas (Arena T2I, AA Image Arena)",
   "tier": 3,
   "status": "active",
   "status_note": "Active",
   "what": "Text-to-image human preference",
   "urls": [
    {
     "url": "https://arena.ai/leaderboard/text-to-image?rankBy=",
     "label": "arena.ai/leaderboard/text-to-image?rankBy=",
     "third_party": false
    },
    {
     "url": "https://artificialanalysis.ai/image/leaderboard/text-to-image",
     "label": "artificialanalysis.ai/image/leaderboard/text-to-image",
     "third_party": false
    }
   ],
   "categories": [
    "Design, UI & multimodal"
   ],
   "category": "Design, UI & multimodal"
  },
  {
   "id": "kaggle-game-arena",
   "name": "Kaggle Game Arena",
   "tier": 3,
   "status": "active",
   "status_note": "Active",
   "what": "Strategic reasoning through chess, poker and Werewolf",
   "urls": [
    {
     "url": "https://www.kaggle.com/benchmarks/kaggle/game-arena",
     "label": "kaggle.com/benchmarks/kaggle/game-arena",
     "third_party": false
    }
   ],
   "categories": [
    "General intelligence & reasoning"
   ],
   "category": "General intelligence & reasoning"
  },
  {
   "id": "ifeval-multichallenge",
   "name": "IFEval / MultiChallenge",
   "tier": 3,
   "status": "saturated",
   "status_note": "Saturated / Active",
   "what": "Verifiable formatting / multi-turn instruction retention",
   "urls": [
    {
     "url": "https://www.llmboard.ai/benchmarks/ifeval-270",
     "label": "llmboard.ai/benchmarks/ifeval-270",
     "third_party": true
    },
    {
     "url": "https://www.llmboard.ai/benchmarks/multichallenge-428",
     "label": "llmboard.ai/benchmarks/multichallenge-428",
     "third_party": true
    }
   ],
   "categories": [
    "Instruction following"
   ],
   "category": "Instruction following"
  },
  {
   "id": "longbench-v2",
   "name": "LongBench v2",
   "tier": 3,
   "status": "active",
   "status_note": "Active",
   "what": "Long-context reasoning",
   "urls": [
    {
     "url": "https://benchlm.ai/benchmarks/longbench-v2",
     "label": "benchlm.ai/benchmarks/longbench-v2",
     "third_party": true
    }
   ],
   "categories": [
    "Long context"
   ],
   "category": "Long context"
  },
  {
   "id": "mmlu-prox-global-mmlu-global-piqa",
   "name": "MMLU-ProX / Global-MMLU / Global PIQA",
   "tier": 3,
   "status": "active",
   "status_note": "Active",
   "what": "Multilingual reasoning (29 / 42 / 100+ languages)",
   "urls": [
    {
     "url": "https://mmluprox.github.io/",
     "label": "mmluprox.github.io",
     "third_party": false
    },
    {
     "url": "https://aclanthology.org/2025.acl-long.919.pdf",
     "label": "aclanthology.org/2025.acl-long.919.pdf",
     "third_party": false
    },
    {
     "url": "https://arxiv.org/html/2510.24081v2",
     "label": "arxiv.org/html/2510.24081v2",
     "third_party": false
    }
   ],
   "categories": [
    "Multilingual (Swedish/Nordic)"
   ],
   "category": "Multilingual (Swedish/Nordic)"
  },
  {
   "id": "helm-capabilities",
   "name": "HELM Capabilities",
   "tier": 3,
   "status": "saturated",
   "status_note": "Mostly saturated components",
   "what": "Stanford holistic suite",
   "urls": [
    {
     "url": "https://www.benchleader.com/benchmarks/helm_mean",
     "label": "benchleader.com/benchmarks/helm_mean",
     "third_party": true
    }
   ],
   "categories": [
    "Aggregators"
   ],
   "category": "Aggregators"
  }
 ],
 "link_summary": {
  "checked": 82,
  "ok": 76,
  "blocked": 5,
  "missing": 0,
  "error": 1,
  "robots": 0,
  "changed": 34
 },
 "link_checks": {
  "https://artificialanalysis.ai/methodology/intelligence-benchmarking": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": true,
   "last_changed": "2026-11-17T05:51:19Z"
  },
  "https://artificialanalysis.ai/evaluations/gdpval-aa?models=": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": true,
   "last_changed": "2026-11-17T05:51:19Z"
  },
  "https://artificialanalysis.ai/evaluations/ifbench": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": true,
   "last_changed": "2026-11-17T05:51:19Z"
  },
  "https://artificialanalysis.ai/image/leaderboard/text-to-image": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": true,
   "last_changed": "2026-11-17T05:51:19Z"
  },
  "https://epoch.ai/eci": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": true,
   "last_changed": "2026-11-17T05:51:19Z"
  },
  "https://epoch.ai/benchmarks": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": true,
   "last_changed": "2026-11-17T05:51:19Z"
  },
  "https://epoch.ai/frontiermath/tiers-1-4": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false,
   "last_changed": "2026-10-27T05:51:01Z"
  },
  "https://epoch.ai/benchmarks/frontiermath-tier-4-v2": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": true,
   "last_changed": "2026-11-17T05:51:19Z"
  },
  "https://epoch.ai/frontiermath/open-problems": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false,
   "last_changed": "2026-10-27T05:51:01Z"
  },
  "https://lastexam.ai/": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false,
   "last_changed": "2026-10-15T04:44:37Z"
  },
  "https://labs.scale.com/leaderboard/humanitys_last_exam": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false,
   "last_changed": "2026-11-09T05:49:50Z"
  },
  "https://labs.scale.com/leaderboard/swe_bench_pro_public": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": true,
   "last_changed": "2026-11-17T05:51:19Z"
  },
  "https://labs.scale.com/leaderboard/rli": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false,
   "last_changed": "2026-11-13T05:52:09Z"
  },
  "https://arcprize.org/leaderboard": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": true,
   "last_changed": "2026-11-17T05:51:19Z"
  },
  "https://arcprize.org/blog/arc-agi-3-launch": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": true,
   "last_changed": "2026-11-17T05:51:19Z"
  },
  "https://www.tbench.ai/news": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": true,
   "last_changed": "2026-11-17T05:51:19Z"
  },
  "https://www.tbench.ai/leaderboard/terminal-bench/2.0": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": true,
   "last_changed": "2026-11-17T05:51:19Z"
  },
  "https://openai.com/index/gdpval/": {
   "status": "blocked",
   "http": 403,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false,
   "last_changed": "2026-11-11T05:48:54Z"
  },
  "https://osworld-v2.xlang.ai/": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false,
   "last_changed": "2026-11-14T05:51:11Z"
  },
  "https://github.com/xlang-ai/OSWorld-V2": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": true,
   "last_changed": "2026-11-17T05:51:19Z"
  },
  "https://github.com/EnvCommons/IMO-Bench": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": true,
   "last_changed": "2026-11-17T05:51:19Z"
  },
  "https://github.com/ServiceNow/webarena-verified": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": true,
   "last_changed": "2026-11-17T05:51:19Z"
  },
  "https://github.com/openai/mle-bench/": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": true,
   "last_changed": "2026-11-17T05:51:19Z"
  },
  "https://github.com/SEC-bench/SEC-bench-Pro": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": true,
   "last_changed": "2026-11-17T05:51:19Z"
  },
  "https://github.com/maveryn/cti-bench": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": true,
   "last_changed": "2026-11-17T05:51:19Z"
  },
  "https://lmarena.ai/leaderboard": {
   "status": "blocked",
   "http": 403,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false,
   "last_changed": "2026-10-13T04:44:09Z"
  },
  "https://lmarena.ai/leaderboard/code": {
   "status": "blocked",
   "http": 403,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false
  },
  "https://arena.ai/blog/lmarena-is-now-arena/": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": true,
   "last_changed": "2026-11-17T05:51:19Z"
  },
  "https://arena.ai/leaderboard/text-to-image?rankBy=": {
   "status": "blocked",
   "http": 403,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false,
   "last_changed": "2026-10-13T04:44:09Z"
  },
  "https://taubench.com/leaderboard": {
   "status": "error",
   "http": 500,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false
  },
  "https://www.remotelabor.ai/": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false,
   "last_changed": "2026-11-13T05:52:09Z"
  },
  "https://andonlabs.com/evals/vending-bench-2": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false,
   "last_changed": "2026-11-14T05:51:11Z"
  },
  "https://metr.org/time-horizons/": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": true,
   "last_changed": "2026-11-17T05:51:19Z"
  },
  "https://www.aisi.gov.uk/blog/how-do-frontier-ai-agents-perform-in-multi-step-cyber-attack-scenarios": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false,
   "last_changed": "2026-11-16T05:49:36Z"
  },
  "https://www.cybergym.io/cybergym/": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": true,
   "last_changed": "2026-11-17T05:51:19Z"
  },
  "https://cybench.github.io/index.html": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false
  },
  "https://atlas.kevinhu.io/benchmarks/livecodebench-pro": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": true,
   "last_changed": "2026-11-17T05:51:19Z"
  },
  "https://texttron.github.io/BrowseComp-Plus/": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false
  },
  "https://benchmarklist.com/benchmarks/mmmu_pro/": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false,
   "last_changed": "2026-11-11T05:48:54Z"
  },
  "https://benchmarklist.com/benchmarks/agentharm/": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false,
   "last_changed": "2026-11-16T05:49:36Z"
  },
  "https://benchmarklist.com/benchmarks/helm_safety/": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false,
   "last_changed": "2026-11-16T05:49:36Z"
  },
  "https://benchmarklist.com/benchmarks/poisonforge_task_level_targeted_poisoning_benchmark_for_instruction_tuned_llms/": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false,
   "last_changed": "2026-11-11T05:48:54Z"
  },
  "https://gui-agent.github.io/grounding-leaderboard/index.html": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false,
   "last_changed": "2026-11-11T05:48:54Z"
  },
  "https://www.designarena.ai/leaderboard": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": true,
   "last_changed": "2026-11-17T05:51:19Z"
  },
  "https://www.kaggle.com/benchmarks/deepmind/simpleqa-verified": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false
  },
  "https://www.kaggle.com/benchmarks/google/facts": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false
  },
  "https://www.kaggle.com/benchmarks/kaggle/game-arena": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false
  },
  "https://huggingface.co/datasets/openai/mrcr": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": true,
   "last_changed": "2026-11-17T05:51:19Z"
  },
  "https://huggingface.co/datasets/ArtificialAnalysis/AA-LCR": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": true,
   "last_changed": "2026-11-17T05:51:19Z"
  },
  "https://matharena.ai/no_final_answer/": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false,
   "last_changed": "2026-11-14T05:51:11Z"
  },
  "https://matharena.ai/competitions": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false,
   "last_changed": "2026-11-14T05:51:11Z"
  },
  "https://critpt.com/": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false,
   "last_changed": "2026-11-14T05:51:11Z"
  },
  "https://arxiv.org/abs/2601.14242": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false,
   "last_changed": "2026-10-31T05:46:41Z"
  },
  "https://arxiv.org/abs/2311.12022": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false,
   "last_changed": "2026-10-31T05:46:41Z"
  },
  "https://arxiv.org/html/2406.01574v5": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false
  },
  "https://arxiv.org/html/2505.15216v3": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false
  },
  "https://arxiv.org/html/2510.24081v2": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false
  },
  "https://livebench.ai/": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": true,
   "last_changed": "2026-11-17T05:51:19Z"
  },
  "https://scale.com/blog/leaderboard": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false,
   "last_changed": "2026-10-16T04:47:44Z"
  },
  "https://euroeval.com/leaderboards/swedish": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false,
   "last_changed": "2026-11-16T05:49:36Z"
  },
  "https://euroeval.com/leaderboards/scandinavian": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false,
   "last_changed": "2026-11-16T05:49:36Z"
  },
  "https://www.swebench.com/": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false
  },
  "https://aider.chat/docs/leaderboards/": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false
  },
  "https://personalagentbench.com/benchmarks/gaia/": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false,
   "last_changed": "2026-11-13T05:52:09Z"
  },
  "https://leaderboard.steel.dev/leaderboards/browsecomp/": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false,
   "last_changed": "2026-11-14T05:51:11Z"
  },
  "https://swelancer.github.io/leaderboard/": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false
  },
  "https://benchlm.ai/benchmarks/paperbench": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": true,
   "last_changed": "2026-11-17T05:51:19Z"
  },
  "https://benchlm.ai/benchmarks/longbench-v2": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": true,
   "last_changed": "2026-11-17T05:51:19Z"
  },
  "https://scicode-bench.github.io/": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false
  },
  "https://zapier.com/benchmarks": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": true,
   "last_changed": "2026-11-17T05:51:19Z"
  },
  "https://meta-llama.github.io/PurpleLlama/CyberSecEval/docs/intro": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false
  },
  "https://dl.acm.org/doi/abs/10.5555/3780338.3783551": {
   "status": "blocked",
   "http": 403,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false
  },
  "https://www.wmdp.ai/": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false,
   "last_changed": "2026-11-14T05:51:11Z"
  },
  "https://mlcommons.org/ailuminate/safety/": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": true,
   "last_changed": "2026-11-17T05:51:19Z"
  },
  "https://bboylyg.github.io/backdoorllm-website.github.io/": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false
  },
  "https://video-mme.github.io/home%5Fpage.html": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false
  },
  "https://www.alphaxiv.org/abs/2604.05015": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": true,
   "last_changed": "2026-11-17T05:51:19Z"
  },
  "https://www.llmboard.ai/benchmarks/ifeval-270": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": true,
   "last_changed": "2026-11-17T05:51:19Z"
  },
  "https://www.llmboard.ai/benchmarks/multichallenge-428": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": true,
   "last_changed": "2026-11-17T05:51:19Z"
  },
  "https://mmluprox.github.io/": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false
  },
  "https://aclanthology.org/2025.acl-long.919.pdf": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": false
  },
  "https://www.benchleader.com/benchmarks/helm_mean": {
   "status": "ok",
   "http": 200,
   "checked": "2026-11-17T05:51:19Z",
   "changed": true,
   "last_changed": "2026-11-17T05:51:19Z"
  }
 },
 "changes": [
  {
   "tag": "Page updated",
   "text": "Artificial Analysis Intelligence Index v4.3.2: the public page changed since the last check.",
   "severity": "info"
  },
  {
   "tag": "Page updated",
   "text": "GDPval / GDPval-AA v2.1: the public page changed since the last check.",
   "severity": "info"
  },
  {
   "tag": "Page updated",
   "text": "IFBench: the public page changed since the last check.",
   "severity": "info"
  },
  {
   "tag": "Page updated",
   "text": "Image arenas (Arena T2I, AA Image Arena): the public page changed since the last check.",
   "severity": "info"
  },
  {
   "tag": "Page updated",
   "text": "Epoch Capabilities Index (ECI) + Benchmarking Hub: the public page changed since the last check.",
   "severity": "info"
  },
  {
   "tag": "Page updated",
   "text": "Epoch Capabilities Index (ECI) + Benchmarking Hub: the public page changed since the last check.",
   "severity": "info"
  },
  {
   "tag": "Page updated",
   "text": "FrontierMath Tiers 1–4 (v2): the public page changed since the last check.",
   "severity": "info"
  },
  {
   "tag": "Page updated",
   "text": "SWE-bench Pro (Scale): the public page changed since the last check.",
   "severity": "info"
  },
  {
   "tag": "Page updated",
   "text": "ARC-AGI-3 (with ARC-AGI-2): the public page changed since the last check.",
   "severity": "info"
  },
  {
   "tag": "Page updated",
   "text": "ARC-AGI-3 (with ARC-AGI-2): the public page changed since the last check.",
   "severity": "info"
  },
  {
   "tag": "Page updated",
   "text": "Terminal-Bench 4.0 (and 3.0): the public page changed since the last check.",
   "severity": "info"
  },
  {
   "tag": "Page updated",
   "text": "Terminal-Bench 4.0 (and 3.0): the public page changed since the last check.",
   "severity": "info"
  },
  {
   "tag": "Page updated",
   "text": "OSWorld 2.0 (and OSWorld-Verified): the public page changed since the last check.",
   "severity": "info"
  },
  {
   "tag": "Page updated",
   "text": "IMO-ProofBench / MathArena proofs: the public page changed since the last check.",
   "severity": "info"
  },
  {
   "tag": "Page updated",
   "text": "WebArena-Verified: the public page changed since the last check.",
   "severity": "info"
  },
  {
   "tag": "Page updated",
   "text": "MLE-bench: the public page changed since the last check.",
   "severity": "info"
  },
  {
   "tag": "Page updated",
   "text": "SEC-bench / SEC-bench Pro: the public page changed since the last check.",
   "severity": "info"
  },
  {
   "tag": "Page updated",
   "text": "CTIBench: the public page changed since the last check.",
   "severity": "info"
  },
  {
   "tag": "Page updated",
   "text": "Arena (formerly LMArena) Text/WebDev/Vision: the public page changed since the last check.",
   "severity": "info"
  },
  {
   "tag": "Page updated",
   "text": "METR Time Horizon 1.1: the public page changed since the last check.",
   "severity": "info"
  },
  {
   "tag": "Page updated",
   "text": "CyberGym: the public page changed since the last check.",
   "severity": "info"
  },
  {
   "tag": "Page updated",
   "text": "LiveCodeBench Pro: the public page changed since the last check.",
   "severity": "info"
  },
  {
   "tag": "Page updated",
   "text": "Design Arena: the public page changed since the last check.",
   "severity": "info"
  },
  {
   "tag": "Page updated",
   "text": "MRCR v2 / AA-LCR v1.1: the public page changed since the last check.",
   "severity": "info"
  },
  {
   "tag": "Page updated",
   "text": "MRCR v2 / AA-LCR v1.1: the public page changed since the last check.",
   "severity": "info"
  },
  {
   "tag": "Page updated",
   "text": "LiveBench: the public page changed since the last check.",
   "severity": "info"
  },
  {
   "tag": "Page updated",
   "text": "PaperBench: the public page changed since the last check.",
   "severity": "info"
  },
  {
   "tag": "Page updated",
   "text": "LongBench v2: the public page changed since the last check.",
   "severity": "info"
  },
  {
   "tag": "Page updated",
   "text": "AutomationBench (Zapier): the public page changed since the last check.",
   "severity": "info"
  },
  {
   "tag": "Page updated",
   "text": "MLCommons AILuminate: the public page changed since the last check.",
   "severity": "info"
  }
 ],
 "health": {
  "sources_total": 4,
  "sources_ok": 4,
  "sources_stale": 0,
  "sources_failed": 0,
  "leaderboards": 19,
  "leaderboards_stale": 0,
  "catalogue_warnings": 5,
  "warnings": 0
 }
}
