{
  "api_version": "1",
  "updated": "2026-10-26",
  "latest_event": "2026-10-26",
  "rights": "© Frontier Log, all rights reserved.",
  "model": {
    "slug": "openai-o4-mini",
    "name": "o4-mini",
    "lab": "openai",
    "date": "2025-04-16",
    "source": "https://openai.com/index/introducing-o3-and-o4-mini",
    "source_title": "Introducing OpenAI o3 and o4-mini",
    "flagship": false,
    "event": "https://frontierlog.ai/event/2025-04-16-o3-and-o4-mini",
    "releases": [],
    "variant_of": null,
    "made_by": "OpenAI",
    "url": "https://frontierlog.ai/model/openai-o4-mini",
    "results": [
      {
        "id": "openai-o4-mini:aime-2024:no-tools:1",
        "model": "openai-o4-mini",
        "model_name": "o4-mini",
        "model_date": "2025-04-16",
        "event": "https://frontierlog.ai/event/2025-04-16-o3-and-o4-mini",
        "benchmark": "aime-2024",
        "benchmark_name": "AIME 2024",
        "setting": "no-tools",
        "setting_name": "No tools",
        "score": 93.4,
        "unit": "%",
        "printed": "93.4",
        "reported_by": "openai",
        "run_by": null,
        "label": "o4-mini (no tools)",
        "source": "https://openai.com/index/introducing-o3-and-o4-mini",
        "source_title": "Introducing OpenAI o3 and o4-mini",
        "source_date": "2025-04-16",
        "where": "'What's changed' section, bar chart 'AIME 2024 / Competition Math', 4th bar 'o4-mini (no tools)' (page screenshot (job 20879 capture), part 02)",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "high reasoning effort",
        "caveat": "",
        "human_verified": false,
        "human_verified_on": null,
        "check": "corroborated",
        "corroborated": true,
        "corroborated_on": "2026-10-26",
        "corroborated_by": [
          {
            "url": "https://blog.google/products-and-platforms/products/gemini/gemini-2-5-flash-preview/",
            "name": "Gemini 2.5 Flash",
            "title": "Developers can now start building with Gemini 2.5 Flash.",
            "lab": "google",
            "date": "2025-04-17",
            "printed": "93.4%",
            "where": "Benchmark table image, row 'Mathematics AIME 2024', single attempt (pass@1), column 'OpenAI o4-mini'; Google (another lab) restating; Google's methodology says non-Gemini results are providers' self-reported numbers, falling back to leaderboards (HLE from agi.safe.ai/Scale, AIME 2025 from matharena) where provider numbers are unavailable"
          }
        ],
        "source_capture": "https://frontierlog.ai/captures/a94b4ebbf6b0463c.png"
      },
      {
        "id": "openai-o4-mini:aime-2025:no-tools:1",
        "model": "openai-o4-mini",
        "model_name": "o4-mini",
        "model_date": "2025-04-16",
        "event": "https://frontierlog.ai/event/2025-04-16-o3-and-o4-mini",
        "benchmark": "aime-2025",
        "benchmark_name": "AIME 2025",
        "setting": "no-tools",
        "setting_name": "No tools",
        "score": 92.7,
        "unit": "%",
        "printed": "92.7",
        "reported_by": "openai",
        "run_by": null,
        "label": "o4-mini (no tools)",
        "source": "https://openai.com/index/introducing-o3-and-o4-mini",
        "source_title": "Introducing OpenAI o3 and o4-mini",
        "source_date": "2025-04-16",
        "where": "'What's changed' section, bar chart 'AIME 2025 / Competition Math', 4th bar 'o4-mini (no tools)' (page screenshot (job 20879 capture), part 02)",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "high reasoning effort",
        "caveat": "",
        "human_verified": false,
        "human_verified_on": null,
        "check": "corroborated",
        "corroborated": true,
        "corroborated_on": "2026-10-26",
        "corroborated_by": [
          {
            "url": "https://blog.google/products-and-platforms/products/gemini/gemini-2-5-flash-preview/",
            "name": "Gemini 2.5 Flash",
            "title": "Developers can now start building with Gemini 2.5 Flash.",
            "lab": "google",
            "date": "2025-04-17",
            "printed": "92.7%",
            "where": "Benchmark table image, row 'Mathematics AIME 2025', single attempt (pass@1), column 'OpenAI o4-mini'; Google (another lab) restating; Google's methodology says non-Gemini results are providers' self-reported numbers, falling back to leaderboards (HLE from agi.safe.ai/Scale, AIME 2025 from matharena) where provider numbers are unavailable"
          },
          {
            "url": "https://blog.google/innovation-and-ai/models-and-research/google-deepmind/google-gemini-updates-io-2025/",
            "name": "Google introduces Deep Think for Gemini 2.5 Pro",
            "title": "Gemini 2.5: Our most intelligent models are getting even better",
            "lab": "google",
            "date": "2025-05-20",
            "printed": "92.7%",
            "where": "Gemini 2.5 Flash (05-20) benchmark table image, row 'Mathematics AIME 2025', single attempt (pass@1), column 'OpenAI o4-mini'; Google restating, same methodology note (AIME 2025 from matharena where provider numbers unavailable)"
          },
          {
            "url": "https://storage.googleapis.com/deepmind-media/gemini/gemini_v2_5_report.pdf",
            "name": "Gemini 2.5 GA and 2.5 Flash-Lite card",
            "title": "Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities",
            "lab": "google",
            "date": "2025-06-17",
            "printed": "92.7%",
            "where": "Gemini 2.5 technical report, Table 4, row 'Math AIME 2025 single attempt', column 'o4-mini high'; Google restating (non-Gemini results from providers' self-reported numbers unless noted)"
          }
        ],
        "source_capture": "https://frontierlog.ai/captures/a94b4ebbf6b0463c.png"
      },
      {
        "id": "openai-o4-mini:gpqa-diamond:default:1",
        "model": "openai-o4-mini",
        "model_name": "o4-mini",
        "model_date": "2025-04-16",
        "event": "https://frontierlog.ai/event/2025-04-16-o3-and-o4-mini",
        "benchmark": "gpqa-diamond",
        "benchmark_name": "GPQA Diamond",
        "setting": "default",
        "setting_name": "",
        "score": 81.4,
        "unit": "%",
        "printed": "81.4",
        "reported_by": "openai",
        "run_by": null,
        "label": "o4-mini (no tools)",
        "source": "https://openai.com/index/introducing-o3-and-o4-mini",
        "source_title": "Introducing OpenAI o3 and o4-mini",
        "source_date": "2025-04-16",
        "where": "'What's changed' section, bar chart 'GPQA Diamond / PhD-Level Science Questions', 4th bar 'o4-mini (no tools)' (page screenshot (job 20879 capture), part 02)",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "high reasoning effort; no tools",
        "caveat": "",
        "human_verified": false,
        "human_verified_on": null,
        "check": "corroborated",
        "corroborated": true,
        "corroborated_on": "2026-10-26",
        "corroborated_by": [
          {
            "url": "https://blog.google/products-and-platforms/products/gemini/gemini-2-5-flash-preview/",
            "name": "Gemini 2.5 Flash",
            "title": "Developers can now start building with Gemini 2.5 Flash.",
            "lab": "google",
            "date": "2025-04-17",
            "printed": "81.4%",
            "where": "Benchmark table image, row 'Science GPQA diamond', single attempt (pass@1), column 'OpenAI o4-mini'; Google restating"
          },
          {
            "url": "https://blog.google/innovation-and-ai/models-and-research/google-deepmind/google-gemini-updates-io-2025/",
            "name": "Google introduces Deep Think for Gemini 2.5 Pro",
            "title": "Gemini 2.5: Our most intelligent models are getting even better",
            "lab": "google",
            "date": "2025-05-20",
            "printed": "81.4%",
            "where": "Gemini 2.5 Flash (05-20) benchmark table image, row 'Science GPQA diamond', single attempt (pass@1), column 'OpenAI o4-mini'; Google restating"
          },
          {
            "url": "https://storage.googleapis.com/deepmind-media/gemini/gemini_v2_5_report.pdf",
            "name": "Gemini 2.5 GA and 2.5 Flash-Lite card",
            "title": "Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities",
            "lab": "google",
            "date": "2025-06-17",
            "printed": "81.4%",
            "where": "Gemini 2.5 technical report, Table 4, row 'Reasoning GPQA (diamond) single attempt', column 'o4-mini high'; Google restating"
          }
        ],
        "source_capture": "https://frontierlog.ai/captures/392f30ae76e5713f.png"
      },
      {
        "id": "openai-o4-mini:hle:no-tools:1",
        "model": "openai-o4-mini",
        "model_name": "o4-mini",
        "model_date": "2025-04-16",
        "event": "https://frontierlog.ai/event/2025-04-16-o3-and-o4-mini",
        "benchmark": "hle",
        "benchmark_name": "Humanity's Last Exam",
        "setting": "no-tools",
        "setting_name": "No tools",
        "score": 14.28,
        "unit": "%",
        "printed": "14.28",
        "reported_by": "openai",
        "run_by": null,
        "label": "o4-mini (no tools)",
        "source": "https://openai.com/index/introducing-o3-and-o4-mini",
        "source_title": "Introducing OpenAI o3 and o4-mini",
        "source_date": "2025-04-16",
        "where": "'What's changed' section, bar chart 'Humanity's Last Exam', 5th bar 'o4-mini (no tools)' (page screenshot (job 20879 capture), part 02)",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "high reasoning effort",
        "caveat": "",
        "human_verified": false,
        "human_verified_on": null,
        "check": "corroborated",
        "corroborated": true,
        "corroborated_on": "2026-10-26",
        "corroborated_by": [
          {
            "url": "https://blog.google/products-and-platforms/products/gemini/gemini-2-5-flash-preview/",
            "name": "Gemini 2.5 Flash",
            "title": "Developers can now start building with Gemini 2.5 Flash.",
            "lab": "google",
            "date": "2025-04-17",
            "printed": "14.3%",
            "where": "Benchmark table image, row 'Humanity's Last Exam (no tools)', column 'OpenAI o4-mini'; Google restating (HLE may come from agi.safe.ai/Scale leaderboard where provider numbers are unavailable)"
          },
          {
            "url": "https://blog.google/innovation-and-ai/models-and-research/google-deepmind/google-gemini-updates-io-2025/",
            "name": "Google introduces Deep Think for Gemini 2.5 Pro",
            "title": "Gemini 2.5: Our most intelligent models are getting even better",
            "lab": "google",
            "date": "2025-05-20",
            "printed": "14.3%",
            "where": "Gemini 2.5 Flash (05-20) benchmark table image, row 'Humanity's Last Exam (no tools)', column 'OpenAI o4-mini'; Google restating, same HLE sourcing caveat"
          }
        ],
        "source_capture": "https://frontierlog.ai/captures/392f30ae76e5713f.png"
      },
      {
        "id": "openai-o4-mini:hle:with-tools:1",
        "model": "openai-o4-mini",
        "model_name": "o4-mini",
        "model_date": "2025-04-16",
        "event": "https://frontierlog.ai/event/2025-04-16-o3-and-o4-mini",
        "benchmark": "hle",
        "benchmark_name": "Humanity's Last Exam",
        "setting": "with-tools",
        "setting_name": "With tools",
        "score": 17.7,
        "unit": "%",
        "printed": "17.70",
        "reported_by": "openai",
        "run_by": null,
        "label": "o4-mini (with python + browsing** tools)",
        "source": "https://openai.com/index/introducing-o3-and-o4-mini",
        "source_title": "Introducing OpenAI o3 and o4-mini",
        "source_date": "2025-04-16",
        "where": "'What's changed' section, bar chart 'Humanity's Last Exam', 6th bar 'o4-mini (with python + browsing** tools)' (page screenshot (job 20879 capture), part 02)",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "high reasoning effort; Python and browsing; known cheating domains blocked; attempts a monitor flags count as wrong",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-26",
        "check": "human",
        "corroborated": false,
        "corroborated_on": null,
        "corroborated_by": [],
        "source_capture": "https://frontierlog.ai/captures/392f30ae76e5713f.png"
      },
      {
        "id": "openai-o4-mini:swe-bench-verified:default:1",
        "model": "openai-o4-mini",
        "model_name": "o4-mini",
        "model_date": "2025-04-16",
        "event": "https://frontierlog.ai/event/2025-04-16-o3-and-o4-mini",
        "benchmark": "swe-bench-verified",
        "benchmark_name": "SWE-bench Verified",
        "setting": "default",
        "setting_name": "",
        "score": 68.1,
        "unit": "%",
        "printed": "68.1",
        "reported_by": "openai",
        "run_by": null,
        "label": "o4-mini",
        "source": "https://openai.com/index/introducing-o3-and-o4-mini",
        "source_title": "Introducing OpenAI o3 and o4-mini",
        "source_date": "2025-04-16",
        "where": "Coding section, bar chart 'SWE-Bench Verified (n=477)', 4th bar 'o4-mini' (page screenshot (job 20879 capture), part 03)",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "high reasoning effort; 477 of the 500 verified tasks; 256K context",
        "caveat": "",
        "human_verified": false,
        "human_verified_on": null,
        "check": "corroborated",
        "corroborated": true,
        "corroborated_on": "2026-10-26",
        "corroborated_by": [
          {
            "url": "https://blog.google/innovation-and-ai/models-and-research/google-deepmind/google-gemini-updates-io-2025/",
            "name": "Google introduces Deep Think for Gemini 2.5 Pro",
            "title": "Gemini 2.5: Our most intelligent models are getting even better",
            "lab": "google",
            "date": "2025-05-20",
            "printed": "68.1%",
            "where": "Gemini 2.5 Flash (05-20) benchmark table image, row 'Agentic coding SWE-bench Verified', column 'OpenAI o4-mini'; Google restating ('All SWE-bench Verified numbers follow official provider reports')"
          },
          {
            "url": "https://storage.googleapis.com/deepmind-media/gemini/gemini_v2_5_report.pdf",
            "name": "Gemini 2.5 GA and 2.5 Flash-Lite card",
            "title": "Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities",
            "lab": "google",
            "date": "2025-06-17",
            "printed": "68.1%",
            "where": "Gemini 2.5 technical report, Table 4, row 'SWE-bench Verified single attempt', column 'o4-mini high'; Google restating official provider report"
          }
        ],
        "source_capture": "https://frontierlog.ai/captures/b532a1f9f0f42f2d.png"
      },
      {
        "id": "openai-o4-mini:browsecomp:default:1",
        "model": "openai-o4-mini",
        "model_name": "o4-mini",
        "model_date": "2025-04-16",
        "event": "https://frontierlog.ai/event/2025-04-16-o3-and-o4-mini",
        "benchmark": "browsecomp",
        "benchmark_name": "BrowseComp",
        "setting": "default",
        "setting_name": "",
        "score": 28.3,
        "unit": "%",
        "printed": "28.3",
        "reported_by": "openai",
        "run_by": null,
        "label": "o4-mini with python + browsing",
        "source": "https://openai.com/index/introducing-o3-and-o4-mini",
        "source_title": "Introducing OpenAI o3 and o4-mini",
        "source_date": "2025-04-16",
        "where": "'Instruction following and agentic tool use' section, bar chart 'BrowseComp / Agentic Browsing', 3rd bar 'o4-mini with python + browsing**' (page screenshot (job 20879 capture), part 04)",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "high reasoning effort; Python and browsing",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-26",
        "check": "human",
        "corroborated": false,
        "corroborated_on": null,
        "corroborated_by": [],
        "source_capture": "https://frontierlog.ai/captures/dc46417a14f5c000.png"
      },
      {
        "id": "openai-o4-mini:aime-2025:with-tools:1",
        "model": "openai-o4-mini",
        "model_name": "o4-mini",
        "model_date": "2025-04-16",
        "event": "https://frontierlog.ai/event/2025-04-16-o3-and-o4-mini",
        "benchmark": "aime-2025",
        "benchmark_name": "AIME 2025",
        "setting": "with-tools",
        "setting_name": "With tools",
        "score": 99.5,
        "unit": "%",
        "printed": "99.5%",
        "reported_by": "openai",
        "run_by": null,
        "label": "o4-mini",
        "source": "https://openai.com/index/introducing-o3-and-o4-mini",
        "source_title": "Introducing OpenAI o3 and o4-mini",
        "source_date": "2025-04-16",
        "where": "'What's changed' section, paragraph on o4-mini, sentence on AIME 2025 with a Python interpreter",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "Python interpreter; pass@1; OpenAI says not to compare with-tools scores to models without tools",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-26",
        "check": "human",
        "corroborated": false,
        "corroborated_on": null,
        "corroborated_by": []
      },
      {
        "id": "openai-o4-mini:aime-2025:with-tools:2",
        "model": "openai-o4-mini",
        "model_name": "o4-mini",
        "model_date": "2025-04-16",
        "event": "https://frontierlog.ai/event/2025-04-16-o3-and-o4-mini",
        "benchmark": "aime-2025",
        "benchmark_name": "AIME 2025",
        "setting": "with-tools",
        "setting_name": "With tools",
        "score": 100.0,
        "unit": "%",
        "printed": "100%",
        "reported_by": "openai",
        "run_by": null,
        "label": "o4-mini",
        "source": "https://openai.com/index/introducing-o3-and-o4-mini",
        "source_title": "Introducing OpenAI o3 and o4-mini",
        "source_date": "2025-04-16",
        "where": "'What's changed' section, paragraph on o4-mini, sentence on AIME 2025 with a Python interpreter, the figure in the parenthesis",
        "own": true,
        "primary": false,
        "parallel": true,
        "notes": "Python interpreter; majority vote over 8 samples",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-26",
        "check": "human",
        "corroborated": false,
        "corroborated_on": null,
        "corroborated_by": []
      }
    ]
  }
}