{
  "api_version": "1",
  "updated": "2026-10-16",
  "latest_event": "2026-10-15",
  "rights": "© Frontier Log, all rights reserved.",
  "benchmark": {
    "slug": "hle",
    "name": "Humanity's Last Exam",
    "family": "hle",
    "unit": "%",
    "higher_is_better": true,
    "status": "active",
    "benchmark_url": "https://lastexam.ai",
    "settings": [
      {
        "setting": "no-tools",
        "name": "No tools",
        "url": "https://frontierlog.ai/benchmark/hle",
        "results": 17
      },
      {
        "setting": "with-tools",
        "name": "With tools",
        "url": "https://frontierlog.ai/benchmark/hle/with-tools",
        "results": 2
      }
    ]
  },
  "total": 19,
  "results": [
    {
      "id": "claude-sonnet-4-5:hle:no-tools:1",
      "model": "claude-sonnet-4-5",
      "model_name": "Claude Sonnet 4.5",
      "model_date": "2025-09-29",
      "event": "https://frontierlog.ai/event/2025-09-29-claude-sonnet-4-5",
      "benchmark": "hle",
      "setting": "no-tools",
      "score": 17.7,
      "unit": "%",
      "printed": "17.7%",
      "reported_by": "anthropic",
      "label": "Sonnet 4.5",
      "source": "https://www.anthropic.com/claude-opus-4-5-system-card",
      "source_title": "Claude Opus 4.5 System Card",
      "source_date": "2025-11-24",
      "where": "Figure 2.16.A 'Humanity's Last Exam (HLE)', panel 'Without Search', bar 'Sonnet 4.5' (page 32)",
      "own": false,
      "primary": true,
      "parallel": false,
      "notes": "Anthropic notes decontamination improved with Opus 4.5, which may affect this score",
      "human_verified": true,
      "human_verified_on": "2026-10-16"
    },
    {
      "id": "claude-haiku-4-5:hle:no-tools:1",
      "model": "claude-haiku-4-5",
      "model_name": "Claude Haiku 4.5",
      "model_date": "2025-10-15",
      "event": "https://frontierlog.ai/event/2025-10-15-claude-haiku-4-5",
      "benchmark": "hle",
      "setting": "no-tools",
      "score": 10.2,
      "unit": "%",
      "printed": "10.2%",
      "reported_by": "anthropic",
      "label": "Haiku 4.5",
      "source": "https://www.anthropic.com/claude-haiku-5-5",
      "source_title": "Introducing Claude Haiku 5.5",
      "source_date": "2026-10-07",
      "where": "Performance table, row 'Multidisciplinary reasoning, Humanity's Last Exam', column 'Haiku 4.5', 'no tools'",
      "own": false,
      "primary": true,
      "parallel": false,
      "notes": "",
      "human_verified": true,
      "human_verified_on": "2026-10-16"
    },
    {
      "id": "gemini-3-pro:hle:no-tools:1",
      "model": "gemini-3-pro",
      "model_name": "Gemini 3 Pro",
      "model_date": "2025-11-18",
      "event": "https://frontierlog.ai/event/2025-11-18-gemini-3-pro",
      "benchmark": "hle",
      "setting": "no-tools",
      "score": 37.5,
      "unit": "%",
      "printed": "37.5%",
      "reported_by": "google",
      "label": "Gemini 3 Pro",
      "source": "https://blog.google/innovation-and-ai/technology/developers-tools/gemini-3-developers/",
      "source_title": "Start building with Gemini 3",
      "source_date": "2025-11-18",
      "where": "benchmark table (image), row 'Humanity's Last Exam', 'No tools', column 'Gemini 3 Pro'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "",
      "human_verified": true,
      "human_verified_on": "2026-10-16",
      "source_image": "https://frontierlog.ai/sources/3fd7d3190ca58049.gif",
      "source_image_original": "https://storage.googleapis.com/gweb-uniblog-publish-prod/documents/gemini_3_table_final_HLE_Tools_on.gif"
    },
    {
      "id": "claude-opus-4-5:hle:no-tools:1",
      "model": "claude-opus-4-5",
      "model_name": "Claude Opus 4.5",
      "model_date": "2025-11-24",
      "event": "https://frontierlog.ai/event/2025-11-24-claude-opus-4-5",
      "benchmark": "hle",
      "setting": "no-tools",
      "score": 30.8,
      "unit": "%",
      "printed": "30.8%",
      "reported_by": "anthropic",
      "label": "Opus 4.5",
      "source": "https://www.anthropic.com/claude-opus-4-5-system-card",
      "source_title": "Claude Opus 4.5 System Card",
      "source_date": "2025-11-24",
      "where": "Figure 2.16.A 'Humanity's Last Exam (HLE)', panel 'Without Search', bar 'Opus 4.5' (page 32)",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "reasoning only, without tools; Claude Sonnet 4.5 as grader (section 2.16)",
      "human_verified": true,
      "human_verified_on": "2026-10-16"
    },
    {
      "id": "gpt-5-2:hle:no-tools:1",
      "model": "gpt-5-2",
      "model_name": "GPT-5.2",
      "model_date": "2025-12-11",
      "event": "https://frontierlog.ai/event/2025-12-11-gpt-5-2",
      "benchmark": "hle",
      "setting": "no-tools",
      "score": 34.5,
      "unit": "%",
      "printed": "34.5%",
      "reported_by": "openai",
      "label": "GPT‑5.2 Thinking",
      "source": "https://openai.com/index/introducing-gpt-5-2",
      "source_title": "Introducing GPT-5.2",
      "source_date": "2025-12-11",
      "where": "science and math table, row 'HLE (no tools)', column 'GPT-5.2 Thinking'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "",
      "human_verified": true,
      "human_verified_on": "2026-10-16"
    },
    {
      "id": "claude-opus-4-6:hle:no-tools:1",
      "model": "claude-opus-4-6",
      "model_name": "Claude Opus 4.6",
      "model_date": "2026-02-05",
      "event": "https://frontierlog.ai/event/2026-02-05-claude-opus-4-6",
      "benchmark": "hle",
      "setting": "no-tools",
      "score": 40.0,
      "unit": "%",
      "printed": "40.0%",
      "reported_by": "anthropic",
      "label": "Opus 4.6",
      "source": "https://www.anthropic.com/news/claude-opus-4-6",
      "source_title": "Claude Opus 4.6",
      "source_date": "2026-02-05",
      "where": "benchmark table (image), row 'Multidisciplinary reasoning, Humanity's Last Exam', column 'Opus 4.6', 'without tools'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "",
      "human_verified": true,
      "human_verified_on": "2026-10-16",
      "source_image": "https://frontierlog.ai/sources/e6c54904fe1e4f51.png",
      "source_image_original": "https://www-cdn.anthropic.com/images/4zrzovbb/website/f9564dd2f758237bd9dbe775674c4a375aff1e8a-2600x2968.png"
    },
    {
      "id": "gemini-3-1-pro:hle:no-tools:1",
      "model": "gemini-3-1-pro",
      "model_name": "Gemini 3.1 Pro",
      "model_date": "2026-02-19",
      "event": "https://frontierlog.ai/event/2026-02-19-gemini-3-1-pro",
      "benchmark": "hle",
      "setting": "no-tools",
      "score": 44.4,
      "unit": "%",
      "printed": "44.4%",
      "reported_by": "google",
      "label": "Gemini 3.1 Pro, Thinking (High)",
      "source": "https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-1-pro/",
      "source_title": "Gemini 3.1 Pro: A smarter model for your most complex tasks",
      "source_date": "2026-02-19",
      "where": "benchmark table (image), row 'Humanity's Last Exam', 'No tools', column 'Gemini 3.1 Pro, Thinking (High)'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "Thinking (High)",
      "human_verified": true,
      "human_verified_on": "2026-10-16",
      "source_image": "https://frontierlog.ai/sources/6b612deebae92383.gif",
      "source_image_original": "https://storage.googleapis.com/gweb-uniblog-publish-prod/original_images/gemini_3-1-pro__benchmarks.gif"
    },
    {
      "id": "gpt-5-4:hle:no-tools:1",
      "model": "gpt-5-4",
      "model_name": "GPT-5.4",
      "model_date": "2026-03-05",
      "event": "https://frontierlog.ai/event/2026-03-05-gpt-5-4",
      "benchmark": "hle",
      "setting": "no-tools",
      "score": 39.8,
      "unit": "%",
      "printed": "39.8%",
      "reported_by": "openai",
      "label": "GPT‑5.4",
      "source": "https://openai.com/index/introducing-gpt-5-4",
      "source_title": "Introducing GPT-5.4",
      "source_date": "2026-03-05",
      "where": "academic table, row 'Humanity's Last Exam (no tools)', column 'GPT‑5.4'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "",
      "human_verified": true,
      "human_verified_on": "2026-10-16"
    },
    {
      "id": "claude-opus-4-7:hle:no-tools:1",
      "model": "claude-opus-4-7",
      "model_name": "Claude Opus 4.7",
      "model_date": "2026-04-16",
      "event": "https://frontierlog.ai/event/2026-04-16-claude-opus-4-7",
      "benchmark": "hle",
      "setting": "no-tools",
      "score": 46.9,
      "unit": "%",
      "printed": "46.9%",
      "reported_by": "anthropic",
      "label": "Opus 4.7",
      "source": "https://www.anthropic.com/news/claude-opus-4-7",
      "source_title": "Introducing Claude Opus 4.7",
      "source_date": "2026-04-16",
      "where": "benchmark table (image), row 'Multidisciplinary reasoning, Humanity's Last Exam', column 'Opus 4.7', 'no tools'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "",
      "human_verified": true,
      "human_verified_on": "2026-10-16",
      "source_image": "https://frontierlog.ai/sources/527cd462f89b6de4.png",
      "source_image_original": "https://www-cdn.anthropic.com/images/4zrzovbb/website/d434d15757c6abac1122af483617741776d5a114-2600x2638.png"
    },
    {
      "id": "claude-opus-4-8:hle:no-tools:1",
      "model": "claude-opus-4-8",
      "model_name": "Claude Opus 4.8",
      "model_date": "2026-05-28",
      "event": "https://frontierlog.ai/event/2026-05-28-claude-opus-4-8",
      "benchmark": "hle",
      "setting": "no-tools",
      "score": 49.8,
      "unit": "%",
      "printed": "49.8%",
      "reported_by": "anthropic",
      "label": "Opus 4.8",
      "source": "https://www.anthropic.com/news/claude-opus-4-8",
      "source_title": "Introducing Claude Opus 4.8",
      "source_date": "2026-05-28",
      "where": "benchmark table (image), row 'Multidisciplinary reasoning, Humanity's Last Exam', column 'Opus 4.8', 'no tools'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "",
      "human_verified": true,
      "human_verified_on": "2026-10-16",
      "source_image": "https://frontierlog.ai/sources/7ad9b784f2579c8a.png",
      "source_image_original": "https://www-cdn.anthropic.com/images/4zrzovbb/website/a9007019094f217e98cb8261a2765d7646c01708-2600x1392.png"
    },
    {
      "id": "claude-sonnet-5:hle:no-tools:1",
      "model": "claude-sonnet-5",
      "model_name": "Claude Sonnet 5",
      "model_date": "2026-06-30",
      "event": "https://frontierlog.ai/event/2026-06-30-claude-sonnet-5",
      "benchmark": "hle",
      "setting": "no-tools",
      "score": 43.2,
      "unit": "%",
      "printed": "43.2%",
      "reported_by": "anthropic",
      "label": "Sonnet 5",
      "source": "https://www.anthropic.com/news/claude-sonnet-5",
      "source_title": "Introducing Claude Sonnet 5",
      "source_date": "2026-06-30",
      "where": "benchmark table (image), row 'Multidisciplinary reasoning, Humanity's Last Exam', column 'Sonnet 5', 'no tools'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "",
      "human_verified": true,
      "human_verified_on": "2026-10-16",
      "source_image": "https://frontierlog.ai/sources/849d6a908f0d46ac.png",
      "source_image_original": "https://www-cdn.anthropic.com/images/4zrzovbb/website/9941d610909f28a504e16dd5af823df172ec6035-2600x1234.png"
    },
    {
      "id": "claude-sonnet-5:hle:no-tools:2",
      "model": "claude-sonnet-5",
      "model_name": "Claude Sonnet 5",
      "model_date": "2026-06-30",
      "event": "https://frontierlog.ai/event/2026-06-30-claude-sonnet-5",
      "benchmark": "hle",
      "setting": "no-tools",
      "score": 43.1,
      "unit": "%",
      "printed": "43.1",
      "reported_by": "anthropic",
      "label": "Claude Sonnet 5",
      "source": "https://www.anthropic.com/claude-sonnet-5-5-system-card",
      "source_title": "Claude Sonnet 5.5 System Card",
      "source_date": "2026-09-28",
      "where": "Section 8.1 evaluation summary table, row 'Humanity's Last Exam, No tools', column 'Claude Sonnet 5'",
      "own": false,
      "primary": false,
      "parallel": false,
      "notes": "",
      "human_verified": true,
      "human_verified_on": "2026-10-16"
    },
    {
      "id": "claude-opus-5:hle:no-tools:1",
      "model": "claude-opus-5",
      "model_name": "Claude Opus 5",
      "model_date": "2026-07-24",
      "event": "https://frontierlog.ai/event/2026-07-24-claude-opus-5",
      "benchmark": "hle",
      "setting": "no-tools",
      "score": 56.3,
      "unit": "%",
      "printed": "56.3%",
      "reported_by": "anthropic",
      "label": "Opus 5",
      "source": "https://www.anthropic.com/news/claude-opus-5",
      "source_title": "Introducing Claude Opus 5",
      "source_date": "2026-07-24",
      "where": "benchmark table (image), row 'Multidisciplinary reasoning, Humanity's Last Exam', column 'Opus 5', 'no tools'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "",
      "human_verified": true,
      "human_verified_on": "2026-10-16",
      "source_image": "https://frontierlog.ai/sources/924165396ca89130.png",
      "source_image_original": "https://www-cdn.anthropic.com/images/4zrzovbb/website/a8fb4f77a9fe240e6f27f3bdc47a137f3c74a29d-2600x2578.png"
    },
    {
      "id": "claude-opus-5:hle:no-tools:2",
      "model": "claude-opus-5",
      "model_name": "Claude Opus 5",
      "model_date": "2026-07-24",
      "event": "https://frontierlog.ai/event/2026-07-24-claude-opus-5",
      "benchmark": "hle",
      "setting": "no-tools",
      "score": 56.6,
      "unit": "%",
      "printed": "56.6",
      "reported_by": "anthropic",
      "label": "Claude Opus 5",
      "source": "https://www.anthropic.com/claude-opus-5-5-system-card",
      "source_title": "Claude Opus 5.5 System Card",
      "source_date": "2026-09-22",
      "where": "Section 8.1 evaluation summary table, row 'Humanity's Last Exam, No tools', column 'Claude Opus 5'",
      "own": false,
      "primary": false,
      "parallel": false,
      "notes": "",
      "human_verified": true,
      "human_verified_on": "2026-10-16"
    },
    {
      "id": "gpt-6-astra:hle:with-tools:1",
      "model": "gpt-6-astra",
      "model_name": "GPT-6 Astra",
      "model_date": "2026-09-03",
      "event": "https://frontierlog.ai/event/2026-09-03-gpt-6-astra",
      "benchmark": "hle",
      "setting": "with-tools",
      "score": 57.2,
      "unit": "%",
      "printed": "57.2%",
      "reported_by": "openai",
      "label": "GPT‑6 Astra",
      "source": "https://openai.com/index/gpt-6-astra",
      "source_title": "GPT-6 Astra: A new generation of intelligence",
      "source_date": "2026-09-03",
      "where": "Academic table, row 'Humanity's Last Exam (w/ tools)', column 'GPT‑6 Astra'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "",
      "human_verified": true,
      "human_verified_on": "2026-10-16"
    },
    {
      "id": "claude-opus-5-5:hle:no-tools:1",
      "model": "claude-opus-5-5",
      "model_name": "Claude Opus 5.5",
      "model_date": "2026-09-22",
      "event": "https://frontierlog.ai/event/2026-09-22-claude-opus-5-5",
      "benchmark": "hle",
      "setting": "no-tools",
      "score": 64.4,
      "unit": "%",
      "printed": "64.4",
      "reported_by": "anthropic",
      "label": "Claude Opus 5.5",
      "source": "https://www.anthropic.com/claude-opus-5-5-system-card",
      "source_title": "Claude Opus 5.5 System Card",
      "source_date": "2026-09-22",
      "where": "Section 8.1 evaluation summary table, row 'Humanity's Last Exam, No tools', column 'Claude Opus 5.5'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "the launch post reports HLE with tools only",
      "human_verified": true,
      "human_verified_on": "2026-10-16"
    },
    {
      "id": "claude-opus-5-5:hle:with-tools:1",
      "model": "claude-opus-5-5",
      "model_name": "Claude Opus 5.5",
      "model_date": "2026-09-22",
      "event": "https://frontierlog.ai/event/2026-09-22-claude-opus-5-5",
      "benchmark": "hle",
      "setting": "with-tools",
      "score": 67.7,
      "unit": "%",
      "printed": "67.7%",
      "reported_by": "anthropic",
      "label": "Opus 5.5",
      "source": "https://www.anthropic.com/claude-opus-5-5",
      "source_title": "Introducing Claude Opus 5.5",
      "source_date": "2026-09-22",
      "where": "benchmark table, row 'Humanity's Last Exam', 'with tools', column 'Opus 5.5'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "max effort",
      "human_verified": true,
      "human_verified_on": "2026-10-16"
    },
    {
      "id": "claude-sonnet-5-5:hle:no-tools:1",
      "model": "claude-sonnet-5-5",
      "model_name": "Claude Sonnet 5.5",
      "model_date": "2026-09-28",
      "event": "https://frontierlog.ai/event/2026-09-28-claude-sonnet-5-5",
      "benchmark": "hle",
      "setting": "no-tools",
      "score": 56.9,
      "unit": "%",
      "printed": "56.9",
      "reported_by": "anthropic",
      "label": "Claude Sonnet 5.5",
      "source": "https://www.anthropic.com/claude-sonnet-5-5-system-card",
      "source_title": "Claude Sonnet 5.5 System Card",
      "source_date": "2026-09-28",
      "where": "Section 8.1 evaluation summary table, row 'Humanity's Last Exam, No tools', column 'Claude Sonnet 5.5'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "the launch post reports HLE with tools only",
      "human_verified": true,
      "human_verified_on": "2026-10-16"
    },
    {
      "id": "claude-haiku-5-5:hle:no-tools:1",
      "model": "claude-haiku-5-5",
      "model_name": "Claude Haiku 5.5",
      "model_date": "2026-10-07",
      "event": "https://frontierlog.ai/event/2026-10-07-claude-haiku-5-5",
      "benchmark": "hle",
      "setting": "no-tools",
      "score": 45.9,
      "unit": "%",
      "printed": "45.9%",
      "reported_by": "anthropic",
      "label": "Haiku 5.5",
      "source": "https://www.anthropic.com/claude-haiku-5-5",
      "source_title": "Introducing Claude Haiku 5.5",
      "source_date": "2026-10-07",
      "where": "Performance table, row 'Multidisciplinary reasoning, Humanity's Last Exam', column 'Haiku 5.5', 'no tools'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "",
      "human_verified": true,
      "human_verified_on": "2026-10-16"
    }
  ]
}