{
  "api_version": "1",
  "updated": "2026-10-26",
  "latest_event": "2026-10-24",
  "rights": "© Frontier Log, all rights reserved.",
  "model": {
    "slug": "gemini-2-5-pro",
    "name": "Gemini 2.5 Pro",
    "lab": "google",
    "date": "2025-03-25",
    "flagship": true,
    "event": "https://frontierlog.ai/event/2025-03-25-gemini-2-5-pro",
    "releases": [
      "https://frontierlog.ai/event/2025-06-17-gemini-2-5-ga-and-2-5-flash"
    ],
    "variant_of": null,
    "made_by": "Google",
    "url": "https://frontierlog.ai/model/gemini-2-5-pro",
    "results": [
      {
        "id": "gemini-2-5-pro:hle:no-tools:1",
        "model": "gemini-2-5-pro",
        "model_name": "Gemini 2.5 Pro",
        "model_date": "2025-03-25",
        "event": "https://frontierlog.ai/event/2025-03-25-gemini-2-5-pro",
        "benchmark": "hle",
        "benchmark_name": "Humanity's Last Exam",
        "setting": "no-tools",
        "setting_name": "No tools",
        "score": 18.8,
        "unit": "%",
        "printed": "18.8%",
        "reported_by": "google",
        "run_by": null,
        "label": "Gemini 2.5 Pro",
        "source": "https://blog.google/innovation-and-ai/models-and-research/google-deepmind/gemini-model-thinking-updates-march-2025/",
        "source_title": "Gemini 2.5: Our most intelligent AI model",
        "source_date": "2025-03-25",
        "where": "benchmark table (image), row 'Reasoning & knowledge, Humanity's Last Exam (no tools)', column 'Gemini 2.5 Pro Experimental (03-25)'",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "pass@1",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-17",
        "check": "human",
        "corroborated": false,
        "corroborated_on": null,
        "corroborated_by": [],
        "source_image": "https://frontierlog.ai/sources/a8f4323681d1a15d.gif",
        "source_image_original": "https://storage.googleapis.com/gweb-uniblog-publish-prod/original_images/gemini_benchmarks_cropped_light2x_1PPmDuP.gif"
      },
      {
        "id": "gemini-2-5-pro:gpqa-diamond:default:1",
        "model": "gemini-2-5-pro",
        "model_name": "Gemini 2.5 Pro",
        "model_date": "2025-03-25",
        "event": "https://frontierlog.ai/event/2025-03-25-gemini-2-5-pro",
        "benchmark": "gpqa-diamond",
        "benchmark_name": "GPQA Diamond",
        "setting": "default",
        "setting_name": "",
        "score": 84.0,
        "unit": "%",
        "printed": "84.0%",
        "reported_by": "google",
        "run_by": null,
        "label": "Gemini 2.5 Pro",
        "source": "https://blog.google/innovation-and-ai/models-and-research/google-deepmind/gemini-model-thinking-updates-march-2025/",
        "source_title": "Gemini 2.5: Our most intelligent AI model",
        "source_date": "2025-03-25",
        "where": "benchmark table (image), row 'Science, GPQA diamond', 'single attempt (pass@1)', column 'Gemini 2.5 Pro Experimental (03-25)'",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "pass@1",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-17",
        "check": "human",
        "corroborated": false,
        "corroborated_on": null,
        "corroborated_by": [],
        "source_image": "https://frontierlog.ai/sources/a8f4323681d1a15d.gif",
        "source_image_original": "https://storage.googleapis.com/gweb-uniblog-publish-prod/original_images/gemini_benchmarks_cropped_light2x_1PPmDuP.gif"
      },
      {
        "id": "gemini-2-5-pro:swe-bench-verified:default:1",
        "model": "gemini-2-5-pro",
        "model_name": "Gemini 2.5 Pro",
        "model_date": "2025-03-25",
        "event": "https://frontierlog.ai/event/2025-03-25-gemini-2-5-pro",
        "benchmark": "swe-bench-verified",
        "benchmark_name": "SWE-bench Verified",
        "setting": "default",
        "setting_name": "",
        "score": 63.8,
        "unit": "%",
        "printed": "63.8%",
        "reported_by": "google",
        "run_by": null,
        "label": "Gemini 2.5 Pro",
        "source": "https://blog.google/innovation-and-ai/models-and-research/google-deepmind/gemini-model-thinking-updates-march-2025/",
        "source_title": "Gemini 2.5: Our most intelligent AI model",
        "source_date": "2025-03-25",
        "where": "benchmark table (image), row 'Agentic coding, SWE-bench verified', column 'Gemini 2.5 Pro Experimental (03-25)'; setup in the Methodology block under the table, second paragraph",
        "own": true,
        "primary": false,
        "parallel": true,
        "notes": "multiple attempts: Google's scaffold draws several trajectories and the model picks one",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-17",
        "check": "human",
        "corroborated": false,
        "corroborated_on": null,
        "corroborated_by": [],
        "source_image": "https://frontierlog.ai/sources/a8f4323681d1a15d.gif",
        "source_image_original": "https://storage.googleapis.com/gweb-uniblog-publish-prod/original_images/gemini_benchmarks_cropped_light2x_1PPmDuP.gif"
      },
      {
        "id": "gemini-2-5-pro:hle:no-tools:2",
        "model": "gemini-2-5-pro",
        "model_name": "Gemini 2.5 Pro",
        "model_date": "2025-03-25",
        "event": "https://frontierlog.ai/event/2025-03-25-gemini-2-5-pro",
        "benchmark": "hle",
        "benchmark_name": "Humanity's Last Exam",
        "setting": "no-tools",
        "setting_name": "No tools",
        "score": 21.6,
        "unit": "%",
        "printed": "21.6%",
        "reported_by": "google",
        "run_by": null,
        "label": "Gemini 2.5 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 2.5 Pro'",
        "own": false,
        "primary": false,
        "parallel": false,
        "notes": "",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-17",
        "check": "human",
        "corroborated": false,
        "corroborated_on": null,
        "corroborated_by": [],
        "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": "gemini-2-5-pro:gpqa-diamond:default:2",
        "model": "gemini-2-5-pro",
        "model_name": "Gemini 2.5 Pro",
        "model_date": "2025-03-25",
        "event": "https://frontierlog.ai/event/2025-03-25-gemini-2-5-pro",
        "benchmark": "gpqa-diamond",
        "benchmark_name": "GPQA Diamond",
        "setting": "default",
        "setting_name": "",
        "score": 86.4,
        "unit": "%",
        "printed": "86.4%",
        "reported_by": "google",
        "run_by": null,
        "label": "Gemini 2.5 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 'GPQA Diamond', 'No tools', column 'Gemini 2.5 Pro'",
        "own": false,
        "primary": false,
        "parallel": false,
        "notes": "",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-17",
        "check": "human",
        "corroborated": true,
        "corroborated_on": "2026-10-19",
        "corroborated_by": [
          {
            "url": "https://www.anthropic.com/news/claude-opus-4-1",
            "name": "Claude Opus 4.1",
            "title": "Claude Opus 4.1",
            "lab": "anthropic",
            "date": "2025-08-05",
            "printed": "86.4%",
            "where": "Benchmark table image, row 'Graduate-level reasoning GPQA Diamond', column 'Gemini 2.5 Pro'"
          },
          {
            "url": "https://www.anthropic.com/news/claude-sonnet-4-5",
            "name": "Claude Agent SDK",
            "title": "Introducing Claude Sonnet 4.5",
            "lab": "anthropic",
            "date": "2025-09-29",
            "printed": "86.4%",
            "where": "Benchmark table image, row 'Graduate-level reasoning GPQA Diamond', column 'Gemini 2.5 Pro'"
          }
        ],
        "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": "gemini-2-5-pro:arc-agi-2:default:1",
        "model": "gemini-2-5-pro",
        "model_name": "Gemini 2.5 Pro",
        "model_date": "2025-03-25",
        "event": "https://frontierlog.ai/event/2025-03-25-gemini-2-5-pro",
        "benchmark": "arc-agi-2",
        "benchmark_name": "ARC-AGI-2",
        "setting": "default",
        "setting_name": "",
        "score": 4.9,
        "unit": "%",
        "printed": "4.9%",
        "reported_by": "google",
        "run_by": null,
        "label": "Gemini 2.5 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 'ARC-AGI-2', 'ARC Prize Verified', column 'Gemini 2.5 Pro'",
        "own": false,
        "primary": true,
        "parallel": false,
        "notes": "ARC Prize Verified",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-17",
        "check": "human",
        "corroborated": false,
        "corroborated_on": null,
        "corroborated_by": [],
        "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": "gemini-2-5-pro:swe-bench-verified:default:2",
        "model": "gemini-2-5-pro",
        "model_name": "Gemini 2.5 Pro",
        "model_date": "2025-03-25",
        "event": "https://frontierlog.ai/event/2025-03-25-gemini-2-5-pro",
        "benchmark": "swe-bench-verified",
        "benchmark_name": "SWE-bench Verified",
        "setting": "default",
        "setting_name": "",
        "score": 59.6,
        "unit": "%",
        "printed": "59.6%",
        "reported_by": "google",
        "run_by": null,
        "label": "Gemini 2.5 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 'SWE-Bench Verified', 'Single attempt', column 'Gemini 2.5 Pro'",
        "own": false,
        "primary": true,
        "parallel": false,
        "notes": "single attempt",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-17",
        "check": "human",
        "corroborated": false,
        "corroborated_on": null,
        "corroborated_by": [],
        "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": "gemini-2-5-pro:aime-2024:no-tools:1",
        "model": "gemini-2-5-pro",
        "model_name": "Gemini 2.5 Pro",
        "model_date": "2025-03-25",
        "event": "https://frontierlog.ai/event/2025-03-25-gemini-2-5-pro",
        "benchmark": "aime-2024",
        "benchmark_name": "AIME 2024",
        "setting": "no-tools",
        "setting_name": "No tools",
        "score": 92.0,
        "unit": "%",
        "printed": "92.0%",
        "reported_by": "google",
        "run_by": null,
        "label": "Gemini 2.5 Pro Experimental (03-25)",
        "source": "https://blog.google/innovation-and-ai/models-and-research/google-deepmind/gemini-model-thinking-updates-march-2025/",
        "source_title": "Gemini 2.5: Our most intelligent AI model",
        "source_date": "2025-03-25",
        "where": "benchmark table image (gemini_benchmarks_cropped GIF), row 'Mathematics AIME 2024', sub-row 'single attempt (pass@1)', column 'Gemini 2.5 Pro Experimental (03-25)'",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "pass@1, averaged over multiple trials",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-18",
        "check": "human",
        "corroborated": true,
        "corroborated_on": "2026-10-19",
        "corroborated_by": [
          {
            "url": "https://qwen.ai/blog?id=qwen3",
            "name": "Qwen3",
            "title": "Qwen3: Think Deeper, Act Faster",
            "lab": "alibaba",
            "date": "2025-04-29",
            "printed": "92.0",
            "where": "Flagship table image (qwen3-235a22.jpg), row AIME'24, column Gemini2.5-Pro; AIME footnote: average accuracy over 64 samples per query"
          }
        ],
        "source_image": "https://frontierlog.ai/sources/a8f4323681d1a15d.gif",
        "source_image_original": "https://storage.googleapis.com/gweb-uniblog-publish-prod/original_images/gemini_benchmarks_cropped_light2x_1PPmDuP.gif"
      },
      {
        "id": "gemini-2-5-pro:aime-2025:no-tools:1",
        "model": "gemini-2-5-pro",
        "model_name": "Gemini 2.5 Pro",
        "model_date": "2025-03-25",
        "event": "https://frontierlog.ai/event/2025-03-25-gemini-2-5-pro",
        "benchmark": "aime-2025",
        "benchmark_name": "AIME 2025",
        "setting": "no-tools",
        "setting_name": "No tools",
        "score": 86.7,
        "unit": "%",
        "printed": "86.7%",
        "reported_by": "google",
        "run_by": null,
        "label": "Gemini 2.5 Pro Experimental (03-25)",
        "source": "https://blog.google/innovation-and-ai/models-and-research/google-deepmind/gemini-model-thinking-updates-march-2025/",
        "source_title": "Gemini 2.5: Our most intelligent AI model",
        "source_date": "2025-03-25",
        "where": "benchmark table image (gemini_benchmarks_cropped GIF), row 'Mathematics AIME 2025', sub-row 'single attempt (pass@1)', column 'Gemini 2.5 Pro Experimental (03-25)'",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "pass@1, averaged over multiple trials",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-18",
        "check": "human",
        "corroborated": true,
        "corroborated_on": "2026-10-19",
        "corroborated_by": [
          {
            "url": "https://qwen.ai/blog?id=qwen3",
            "name": "Qwen3",
            "title": "Qwen3: Think Deeper, Act Faster",
            "lab": "alibaba",
            "date": "2025-04-29",
            "printed": "86.7",
            "where": "Flagship table image (qwen3-235a22.jpg), row AIME'25, column Gemini2.5-Pro; AIME footnote: average accuracy over 64 samples per query"
          }
        ],
        "source_image": "https://frontierlog.ai/sources/a8f4323681d1a15d.gif",
        "source_image_original": "https://storage.googleapis.com/gweb-uniblog-publish-prod/original_images/gemini_benchmarks_cropped_light2x_1PPmDuP.gif"
      },
      {
        "id": "gemini-2-5-pro:aime-2025:no-tools:2",
        "model": "gemini-2-5-pro",
        "model_name": "Gemini 2.5 Pro",
        "model_date": "2025-03-25",
        "event": "https://frontierlog.ai/event/2025-03-25-gemini-2-5-pro",
        "benchmark": "aime-2025",
        "benchmark_name": "AIME 2025",
        "setting": "no-tools",
        "setting_name": "No tools",
        "score": 88.0,
        "unit": "%",
        "printed": "88.0%",
        "reported_by": "google",
        "run_by": null,
        "label": "Gemini 2.5 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 'AIME 2025', sub-row 'No tools', column 'Gemini 2.5 Pro'",
        "own": false,
        "primary": false,
        "parallel": false,
        "notes": "",
        "caveat": "",
        "human_verified": false,
        "human_verified_on": null,
        "check": "corroborated",
        "corroborated": true,
        "corroborated_on": "2026-10-20",
        "corroborated_by": [
          {
            "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": "88.0%",
            "where": "Gemini 2.5 tech report (linked from the model-family post), comparison table row 'AIME 2025', column 'Gemini 2.5 Pro'; also sentence 'performance on AIME 2025 is 88.0% for Gemini 2.5 Pro'"
          },
          {
            "url": "https://blog.google/products-and-platforms/products/gemini/gemini-3-flash/",
            "name": "Google post",
            "title": "Gemini 3 Flash: frontier intelligence built for speed",
            "lab": "google",
            "date": "2025-12-17",
            "printed": "88.0%",
            "where": "benchmark table image, row 'AIME 2025', sub-row 'No tools', column 'Gemini 2.5 Pro Thinking'"
          },
          {
            "url": "https://blog.google/products-and-platforms/products/gemini/gemini-2-5-pro-latest-preview/",
            "name": "Google post",
            "title": "Try the latest Gemini 2.5 Pro before general availability.",
            "lab": "google",
            "date": "2025-06-05",
            "printed": "88.0%",
            "where": "benchmark table image, row 'Mathematics / AIME 2025', sub-row 'single attempt', column 'Gemini 2.5 Pro Preview 06-05 Thinking' (no tools label printed)"
          },
          {
            "url": "https://blog.google/products-and-platforms/products/gemini/gemini-2-5-model-family-expands/",
            "name": "Gemini 2.5 GA and 2.5 Flash-Lite",
            "title": "We’re expanding our Gemini 2.5 family of models",
            "lab": "google",
            "date": "2025-06-17",
            "printed": "88.0%",
            "where": "benchmark table image, row 'Mathematics / AIME 2025', column 'Gemini 2.5 Pro Thinking' (no tools label printed)"
          }
        ],
        "source_image": "https://frontierlog.ai/sources/3fd7d3190ca58049.gif",
        "source_image_original": "https://storage.googleapis.com/gweb-uniblog-publish-prod/original_images/gemini_3_table_final_HLE_Tools_on_1sXRb4o.gif"
      }
    ]
  }
}