{
  "api_version": "1",
  "updated": "2026-10-26",
  "latest_event": "2026-10-24",
  "rights": "© Frontier Log, all rights reserved.",
  "model": {
    "slug": "gemini-3-1-pro",
    "name": "Gemini 3.1 Pro",
    "lab": "google",
    "date": "2026-02-19",
    "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",
    "flagship": true,
    "event": "https://frontierlog.ai/event/2026-02-19-gemini-3-1-pro",
    "releases": [],
    "variant_of": null,
    "made_by": "Google",
    "url": "https://frontierlog.ai/model/gemini-3-1-pro",
    "results": [
      {
        "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",
        "benchmark_name": "Humanity's Last Exam",
        "setting": "no-tools",
        "setting_name": "No tools",
        "score": 44.4,
        "unit": "%",
        "printed": "44.4%",
        "reported_by": "google",
        "run_by": null,
        "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)",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-16",
        "check": "human",
        "corroborated": true,
        "corroborated_on": "2026-10-19",
        "corroborated_by": [
          {
            "url": "https://www.anthropic.com/claude-fable-5-mythos-5-system-card",
            "name": "Claude Fable 5 and Mythos 5 card",
            "title": "Claude Fable 5 and Claude Mythos 5 System Card",
            "lab": "anthropic",
            "date": "2026-06-09",
            "printed": "44.4",
            "where": "Table 8.1.A (p.251), row Humanity's Last Exam / No tools, column Gemini 3.1 Pro"
          },
          {
            "url": "https://www.anthropic.com/claude-opus-4-7-system-card",
            "name": "Claude Opus 4.7 card",
            "title": "Claude Opus 4.7 System Card",
            "lab": "anthropic",
            "date": "2026-04-16",
            "printed": "44.4%",
            "where": "Table 8.1.A (p.191), row Humanity's Last Exam / No tools, column Gemini 3.1 Pro"
          },
          {
            "url": "https://www.anthropic.com/claude-opus-4-8-system-card",
            "name": "Claude Opus 4.8 card",
            "title": "Claude Opus 4.8 System Card",
            "lab": "anthropic",
            "date": "2026-05-28",
            "printed": "44.4",
            "where": "Table 8.1.A (p.194), row Humanity's Last Exam / No tools, column Gemini 3.1 Pro"
          },
          {
            "url": "https://www.anthropic.com/news/claude-opus-4-7",
            "name": "Claude Opus 4.7",
            "title": "Introducing Claude Opus 4.7",
            "lab": "anthropic",
            "date": "2026-04-16",
            "printed": "44.4%",
            "where": "benchmark table image, row Humanity's Last Exam (no tools), column Gemini 3.1 Pro"
          },
          {
            "url": "https://www.anthropic.com/news/claude-opus-4-8",
            "name": "Claude Opus 4.8",
            "title": "Introducing Claude Opus 4.8",
            "lab": "anthropic",
            "date": "2026-05-28",
            "printed": "44.4%",
            "where": "benchmark table image, row Humanity's Last Exam (no tools), column Gemini 3.1 Pro"
          },
          {
            "url": "https://api-docs.deepseek.com/news/news260424/",
            "name": "DeepSeek V4 Preview",
            "title": "DeepSeek V4 Preview Release",
            "lab": "deepseek",
            "date": "2026-04-24",
            "printed": "44.4",
            "where": "Bar chart image; group 'HLE (Pass@1), under 'Knowledge & Reasoning''; value label on Gemini-3.1-Pro-High bar"
          },
          {
            "url": "https://www.kimi.ai/blog/kimi-k2-6",
            "name": "Kimi K2.6",
            "title": "Kimi K2.6: Advancing Open-Source Coding",
            "lab": "moonshot",
            "date": "2026-04-20",
            "printed": "44.4",
            "where": "Benchmark Table (appendix), row HLE-Full (Reasoning & Knowledge; no tools) ; column Gemini 3.1 Pro (thinking high)"
          },
          {
            "url": "https://openai.com/index/introducing-gpt-5-5",
            "name": "GPT-5.5",
            "title": "Introducing GPT-5.5",
            "lab": "openai",
            "date": "2026-04-23",
            "printed": "44.4%",
            "where": "table 5 \"Humanity's Last Exam (no tools)\" row, \"Gemini 3.1 Pro\" column"
          }
        ],
        "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": "gemini-3-1-pro:arc-agi-2:default: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": "arc-agi-2",
        "benchmark_name": "ARC-AGI-2",
        "setting": "default",
        "setting_name": "",
        "score": 77.1,
        "unit": "%",
        "printed": "77.1%",
        "reported_by": "google",
        "run_by": null,
        "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 'ARC-AGI-2', 'ARC Prize Verified', column 'Gemini 3.1 Pro'",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "ARC Prize Verified; Thinking (High)",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-16",
        "check": "human",
        "corroborated": true,
        "corroborated_on": "2026-10-19",
        "corroborated_by": [
          {
            "url": "https://www.anthropic.com/claude-opus-4-7-system-card",
            "name": "Claude Opus 4.7 card",
            "title": "Claude Opus 4.7 System Card",
            "lab": "anthropic",
            "date": "2026-04-16",
            "printed": "77.1%",
            "where": "Table 8.1.A (p.192), row ARC-AGI-2, column Gemini 3.1 Pro"
          },
          {
            "url": "https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-5/",
            "name": "Gemini 3.5 Flash",
            "title": "Gemini 3.5: frontier intelligence with action",
            "lab": "google",
            "date": "2026-05-19",
            "printed": "77.1%",
            "where": "Benchmark table image; row 'ARC-AGI-2'; column gemini-3-1-pro"
          },
          {
            "url": "https://openai.com/index/introducing-gpt-5-5",
            "name": "GPT-5.5",
            "title": "Introducing GPT-5.5",
            "lab": "openai",
            "date": "2026-04-23",
            "printed": "77.1%",
            "where": "table 8 \"ARC-AGI-2 (Verified)\" row, \"Gemini 3.1 Pro\" column"
          }
        ],
        "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": "gemini-3-1-pro:gpqa-diamond:default: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": "gpqa-diamond",
        "benchmark_name": "GPQA Diamond",
        "setting": "default",
        "setting_name": "",
        "score": 94.3,
        "unit": "%",
        "printed": "94.3%",
        "reported_by": "google",
        "run_by": null,
        "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 'GPQA Diamond', 'No tools', column 'Gemini 3.1 Pro'",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "no tools; Thinking (High)",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-16",
        "check": "human",
        "corroborated": true,
        "corroborated_on": "2026-10-19",
        "corroborated_by": [
          {
            "url": "https://www.anthropic.com/claude-opus-4-7-system-card",
            "name": "Claude Opus 4.7 card",
            "title": "Claude Opus 4.7 System Card",
            "lab": "anthropic",
            "date": "2026-04-16",
            "printed": "94.3%",
            "where": "Table 8.1.A (p.191), row GPQA Diamond, column Gemini 3.1 Pro"
          },
          {
            "url": "https://www.anthropic.com/claude-opus-4-8-system-card",
            "name": "Claude Opus 4.8 card",
            "title": "Claude Opus 4.8 System Card",
            "lab": "anthropic",
            "date": "2026-05-28",
            "printed": "94.3",
            "where": "Table 8.1.A (p.194), row GPQA Diamond, column Gemini 3.1 Pro"
          },
          {
            "url": "https://www.anthropic.com/news/claude-opus-4-7",
            "name": "Claude Opus 4.7",
            "title": "Introducing Claude Opus 4.7",
            "lab": "anthropic",
            "date": "2026-04-16",
            "printed": "94.3%",
            "where": "benchmark table image, row GPQA Diamond, column Gemini 3.1 Pro"
          },
          {
            "url": "https://api-docs.deepseek.com/news/news260424/",
            "name": "DeepSeek V4 Preview",
            "title": "DeepSeek V4 Preview Release",
            "lab": "deepseek",
            "date": "2026-04-24",
            "printed": "94.3",
            "where": "Full benchmark table image; row 'GPQA Diamond (Pass@1)'; column 'Gemini-3.1-Pro High'"
          },
          {
            "url": "https://www.kimi.ai/blog/kimi-k2-6",
            "name": "Kimi K2.6",
            "title": "Kimi K2.6: Advancing Open-Source Coding",
            "lab": "moonshot",
            "date": "2026-04-20",
            "printed": "94.3",
            "where": "Benchmark Table (appendix), row GPQA-Diamond ; column Gemini 3.1 Pro (thinking high)"
          },
          {
            "url": "https://openai.com/index/introducing-gpt-5-5",
            "name": "GPT-5.5",
            "title": "Introducing GPT-5.5",
            "lab": "openai",
            "date": "2026-04-23",
            "printed": "94.3%",
            "where": "table 5 \"GPQA Diamond\" row, \"Gemini 3.1 Pro\" column"
          },
          {
            "url": "https://openai.com/index/gpt-5-6",
            "name": "GPT-5.6",
            "title": "GPT-5.6: Frontier intelligence that scales with your ambition",
            "lab": "openai",
            "date": "2026-07-09",
            "printed": "94.3%",
            "where": "table 7 \"GPQA Diamond\" row, \"Gemini 3.1 Pro Preview\" column"
          },
          {
            "url": "https://z.ai/blog/glm-5.1",
            "name": "GLM-5.1",
            "title": "GLM-5.1: Towards Long-Horizon Tasks",
            "lab": "zai",
            "date": "2026-04-07",
            "printed": "94.3",
            "where": "Benchmark table, row 'GPQA-Diamond', column 'Gemini 3.1 Pro'"
          },
          {
            "url": "https://z.ai/blog/glm-5.2",
            "name": "GLM-5.2",
            "title": "GLM-5.2: Built for Long-Horizon Tasks",
            "lab": "zai",
            "date": "2026-06-16",
            "printed": "94.3",
            "where": "Full Benchmark Table, row 'GPQA-Diamond', column 'Gemini 3.1 Pro'"
          }
        ],
        "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": "gemini-3-1-pro:terminal-bench-2-0:default: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": "terminal-bench-2-0",
        "benchmark_name": "Terminal-Bench 2.0",
        "setting": "default",
        "setting_name": "",
        "score": 68.5,
        "unit": "%",
        "printed": "68.5%",
        "reported_by": "google",
        "run_by": null,
        "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 'Terminal-Bench 2.0', 'Terminus-2 harness', column 'Gemini 3.1 Pro'",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "Terminus 2 harness; Thinking (High)",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-16",
        "check": "human",
        "corroborated": true,
        "corroborated_on": "2026-10-19",
        "corroborated_by": [
          {
            "url": "https://www.anthropic.com/claude-opus-4-7-system-card",
            "name": "Claude Opus 4.7 card",
            "title": "Claude Opus 4.7 System Card",
            "lab": "anthropic",
            "date": "2026-04-16",
            "printed": "68.5%",
            "where": "Table 8.1.A (p.191), row Terminal-Bench 2.0, column Gemini 3.1 Pro"
          },
          {
            "url": "https://www.anthropic.com/news/claude-opus-4-7",
            "name": "Claude Opus 4.7",
            "title": "Introducing Claude Opus 4.7",
            "lab": "anthropic",
            "date": "2026-04-16",
            "printed": "68.5%",
            "where": "benchmark table image, row Terminal-Bench 2.0, column Gemini 3.1 Pro"
          },
          {
            "url": "https://api-docs.deepseek.com/news/news260424/",
            "name": "DeepSeek V4 Preview",
            "title": "DeepSeek V4 Preview Release",
            "lab": "deepseek",
            "date": "2026-04-24",
            "printed": "68.5",
            "where": "Bar chart image; group 'Terminal Bench 2.0 (Acc)'; value label on Gemini-3.1-Pro-High bar"
          },
          {
            "url": "https://www.kimi.ai/blog/kimi-k2-6",
            "name": "Kimi K2.6",
            "title": "Kimi K2.6: Advancing Open-Source Coding",
            "lab": "moonshot",
            "date": "2026-04-20",
            "printed": "68.5",
            "where": "Benchmark Table (appendix), row Terminal-Bench 2.0 (Terminus-2) ; column Gemini 3.1 Pro (thinking high)"
          },
          {
            "url": "https://openai.com/index/introducing-gpt-5-5",
            "name": "GPT-5.5",
            "title": "Introducing GPT-5.5",
            "lab": "openai",
            "date": "2026-04-23",
            "printed": "68.5%",
            "where": "table 0 \"Terminal-Bench 2.0\" row, \"Gemini 3.1 Pro\" column"
          },
          {
            "url": "https://z.ai/blog/glm-5.1",
            "name": "GLM-5.1",
            "title": "GLM-5.1: Towards Long-Horizon Tasks",
            "lab": "zai",
            "date": "2026-04-07",
            "printed": "68.5",
            "where": "Benchmark table, row 'Terminal-Bench 2.0 Terminus-2', column 'Gemini 3.1 Pro'"
          }
        ],
        "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": "gemini-3-1-pro:swe-bench-verified:default: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": "swe-bench-verified",
        "benchmark_name": "SWE-bench Verified",
        "setting": "default",
        "setting_name": "",
        "score": 80.6,
        "unit": "%",
        "printed": "80.6%",
        "reported_by": "google",
        "run_by": null,
        "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 'SWE-Bench Verified', 'Single attempt', column 'Gemini 3.1 Pro'",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "single attempt; Thinking (High)",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-16",
        "check": "human",
        "corroborated": true,
        "corroborated_on": "2026-10-19",
        "corroborated_by": [
          {
            "url": "https://www.anthropic.com/claude-fable-5-mythos-5-system-card",
            "name": "Claude Fable 5 and Mythos 5 card",
            "title": "Claude Fable 5 and Claude Mythos 5 System Card",
            "lab": "anthropic",
            "date": "2026-06-09",
            "printed": "80.6",
            "where": "Table 8.1.A (p.251), row SWE-bench Verified, column Gemini 3.1 Pro"
          },
          {
            "url": "https://www.anthropic.com/claude-opus-4-7-system-card",
            "name": "Claude Opus 4.7 card",
            "title": "Claude Opus 4.7 System Card",
            "lab": "anthropic",
            "date": "2026-04-16",
            "printed": "80.6%",
            "where": "Table 8.1.A (p.191), row SWE-bench Verified, column Gemini 3.1 Pro"
          },
          {
            "url": "https://www.anthropic.com/claude-opus-4-8-system-card",
            "name": "Claude Opus 4.8 card",
            "title": "Claude Opus 4.8 System Card",
            "lab": "anthropic",
            "date": "2026-05-28",
            "printed": "80.6",
            "where": "Table 8.1.A (p.194), row SWE-bench Verified, column Gemini 3.1 Pro"
          },
          {
            "url": "https://www.anthropic.com/news/claude-opus-4-7",
            "name": "Claude Opus 4.7",
            "title": "Introducing Claude Opus 4.7",
            "lab": "anthropic",
            "date": "2026-04-16",
            "printed": "80.6%",
            "where": "benchmark table image, row SWE-bench Verified, column Gemini 3.1 Pro"
          },
          {
            "url": "https://api-docs.deepseek.com/news/news260424/",
            "name": "DeepSeek V4 Preview",
            "title": "DeepSeek V4 Preview Release",
            "lab": "deepseek",
            "date": "2026-04-24",
            "printed": "80.6",
            "where": "Bar chart image; group 'SWE Verified (Resolved)'; value label on Gemini-3.1-Pro-High bar"
          },
          {
            "url": "https://www.minimax.io/blog/minimax-m3",
            "name": "MiniMax M3",
            "title": "MiniMax M3: Frontier Coding, 1M Context, Native Multimodality — All in One Model",
            "lab": "minimax",
            "date": "2026-06-01",
            "printed": "80.6",
            "where": "Full benchmark table image; row 'SWE-Bench Verified'; column 'Gemini 3.1 Pro'"
          },
          {
            "url": "https://www.kimi.ai/blog/kimi-k2-6",
            "name": "Kimi K2.6",
            "title": "Kimi K2.6: Advancing Open-Source Coding",
            "lab": "moonshot",
            "date": "2026-04-20",
            "printed": "80.6",
            "where": "Benchmark Table (appendix), row SWE-Bench Verified ; column Gemini 3.1 Pro (thinking high)"
          },
          {
            "url": "https://thinkingmachines.ai/news/introducing-inkling/",
            "name": "Inkling",
            "title": "Inkling: Our Open-Weights Model",
            "lab": "thinking-machines",
            "date": "2026-07-15",
            "printed": "80.6%",
            "where": "'Benchmarking Inkling' main table, row 'SWEBench Verified* (Inkling bash-only harness; external models self-reported)', column 'Gemini 3.1 Pro (high)'"
          }
        ],
        "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": "gemini-3-1-pro:hle:with-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",
        "benchmark_name": "Humanity's Last Exam",
        "setting": "with-tools",
        "setting_name": "With tools",
        "score": 51.4,
        "unit": "%",
        "printed": "51.4%",
        "reported_by": "google",
        "run_by": null,
        "label": "Gemini 3.1 Pro",
        "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', 'Search (blocklist) + Code', column 'Gemini 3.1 Pro, Thinking (High)'",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "search (with a blocklist) and code execution; Thinking (High)",
        "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/claude-fable-5-mythos-5-system-card",
            "name": "Claude Fable 5 and Mythos 5 card",
            "title": "Claude Fable 5 and Claude Mythos 5 System Card",
            "lab": "anthropic",
            "date": "2026-06-09",
            "printed": "51.4",
            "where": "Table 8.1.A (p.251), row Humanity's Last Exam / With tools, column Gemini 3.1 Pro"
          },
          {
            "url": "https://www.anthropic.com/claude-opus-4-7-system-card",
            "name": "Claude Opus 4.7 card",
            "title": "Claude Opus 4.7 System Card",
            "lab": "anthropic",
            "date": "2026-04-16",
            "printed": "51.4%",
            "where": "Table 8.1.A (p.191), row Humanity's Last Exam / With tools, column Gemini 3.1 Pro"
          },
          {
            "url": "https://www.anthropic.com/claude-opus-4-8-system-card",
            "name": "Claude Opus 4.8 card",
            "title": "Claude Opus 4.8 System Card",
            "lab": "anthropic",
            "date": "2026-05-28",
            "printed": "51.4",
            "where": "Table 8.1.A (p.194), row Humanity's Last Exam / With tools, column Gemini 3.1 Pro"
          },
          {
            "url": "https://www.anthropic.com/news/claude-opus-4-7",
            "name": "Claude Opus 4.7",
            "title": "Introducing Claude Opus 4.7",
            "lab": "anthropic",
            "date": "2026-04-16",
            "printed": "51.4%",
            "where": "benchmark table image, row Humanity's Last Exam (with tools), column Gemini 3.1 Pro"
          },
          {
            "url": "https://www.anthropic.com/news/claude-opus-4-8",
            "name": "Claude Opus 4.8",
            "title": "Introducing Claude Opus 4.8",
            "lab": "anthropic",
            "date": "2026-05-28",
            "printed": "51.4%",
            "where": "benchmark table image, row Humanity's Last Exam (with tools), column Gemini 3.1 Pro"
          },
          {
            "url": "https://www.kimi.ai/blog/kimi-k2-6",
            "name": "Kimi K2.6",
            "title": "Kimi K2.6: Advancing Open-Source Coding",
            "lab": "moonshot",
            "date": "2026-04-20",
            "printed": "51.4",
            "where": "Benchmark Table (appendix), row HLE-Full w/ tools ; column Gemini 3.1 Pro (thinking high)"
          },
          {
            "url": "https://openai.com/index/introducing-gpt-5-5",
            "name": "GPT-5.5",
            "title": "Introducing GPT-5.5",
            "lab": "openai",
            "date": "2026-04-23",
            "printed": "51.4%",
            "where": "table 5 \"Humanity's Last Exam (with tools)\" row, \"Gemini 3.1 Pro\" column"
          },
          {
            "url": "https://thinkingmachines.ai/news/introducing-inkling/",
            "name": "Inkling",
            "title": "Inkling: Our Open-Weights Model",
            "lab": "thinking-machines",
            "date": "2026-07-15",
            "printed": "51.4%",
            "where": "'Benchmarking Inkling' main table, row 'HLE with tools', column 'Gemini 3.1 Pro (high)'"
          },
          {
            "url": "https://z.ai/blog/glm-5.1",
            "name": "GLM-5.1",
            "title": "GLM-5.1: Towards Long-Horizon Tasks",
            "lab": "zai",
            "date": "2026-04-07",
            "printed": "51.4",
            "where": "Benchmark table, row 'HLE w/ Tools (* = full set for Opus 4.6, Gemini 3.1 Pro, GPT-5.4)', column 'Gemini 3.1 Pro'"
          },
          {
            "url": "https://z.ai/blog/glm-5.2",
            "name": "GLM-5.2",
            "title": "GLM-5.2: Built for Long-Horizon Tasks",
            "lab": "zai",
            "date": "2026-06-16",
            "printed": "51.4",
            "where": "Full Benchmark Table, row 'HLE w/ Tools (* = full set for Opus 4.8, GPT-5.5, Gemini 3.1 Pro)', column 'Gemini 3.1 Pro'"
          }
        ],
        "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": "gemini-3-1-pro:browsecomp:default: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": "browsecomp",
        "benchmark_name": "BrowseComp",
        "setting": "default",
        "setting_name": "",
        "score": 85.9,
        "unit": "%",
        "printed": "85.9%",
        "reported_by": "google",
        "run_by": null,
        "label": "Gemini 3.1 Pro",
        "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 'BrowseComp' (Search + Python + Browse), column 'Gemini 3.1 Pro Thinking (High)'",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "search, Python and browsing; Thinking (High)",
        "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/claude-fable-5-mythos-5-system-card",
            "name": "Claude Fable 5 and Mythos 5 card",
            "title": "Claude Fable 5 and Claude Mythos 5 System Card",
            "lab": "anthropic",
            "date": "2026-06-09",
            "printed": "85.9",
            "where": "Table 8.1.A (p.251), row BrowseComp, column Gemini 3.1 Pro"
          },
          {
            "url": "https://www.anthropic.com/claude-opus-4-7-system-card",
            "name": "Claude Opus 4.7 card",
            "title": "Claude Opus 4.7 System Card",
            "lab": "anthropic",
            "date": "2026-04-16",
            "printed": "85.9%",
            "where": "Table 8.1.A (p.191), row BrowseComp, column Gemini 3.1 Pro"
          },
          {
            "url": "https://www.anthropic.com/claude-opus-4-8-system-card",
            "name": "Claude Opus 4.8 card",
            "title": "Claude Opus 4.8 System Card",
            "lab": "anthropic",
            "date": "2026-05-28",
            "printed": "85.9",
            "where": "Table 8.1.A (p.194), row BrowseComp, column Gemini 3.1 Pro"
          },
          {
            "url": "https://www.anthropic.com/news/claude-opus-4-7",
            "name": "Claude Opus 4.7",
            "title": "Introducing Claude Opus 4.7",
            "lab": "anthropic",
            "date": "2026-04-16",
            "printed": "85.9%",
            "where": "benchmark table image, row BrowseComp, column Gemini 3.1 Pro"
          },
          {
            "url": "https://api-docs.deepseek.com/news/news260424/",
            "name": "DeepSeek V4 Preview",
            "title": "DeepSeek V4 Preview Release",
            "lab": "deepseek",
            "date": "2026-04-24",
            "printed": "85.9",
            "where": "Full benchmark table image; row 'BrowseComp (Pass@1)'; column 'Gemini-3.1-Pro High'"
          },
          {
            "url": "https://www.minimax.io/blog/minimax-m3",
            "name": "MiniMax M3",
            "title": "MiniMax M3: Frontier Coding, 1M Context, Native Multimodality — All in One Model",
            "lab": "minimax",
            "date": "2026-06-01",
            "printed": "85.9",
            "where": "Bar chart image (10 panels); panel 'BrowseComp'; value label on Gemini 3.1 Pro bar"
          },
          {
            "url": "https://www.kimi.ai/blog/kimi-k2-6",
            "name": "Kimi K2.6",
            "title": "Kimi K2.6: Advancing Open-Source Coding",
            "lab": "moonshot",
            "date": "2026-04-20",
            "printed": "85.9",
            "where": "Benchmark Table (appendix), row BrowseComp ; column Gemini 3.1 Pro (thinking high)"
          },
          {
            "url": "https://openai.com/index/introducing-gpt-5-5",
            "name": "GPT-5.5",
            "title": "Introducing GPT-5.5",
            "lab": "openai",
            "date": "2026-04-23",
            "printed": "85.9%",
            "where": "table 0 \"BrowseComp\" row, \"Gemini 3.1 Pro\" column"
          },
          {
            "url": "https://openai.com/index/gpt-5-6",
            "name": "GPT-5.6",
            "title": "GPT-5.6: Frontier intelligence that scales with your ambition",
            "lab": "openai",
            "date": "2026-07-09",
            "printed": "85.9%",
            "where": "table 3 \"BrowseComp\" row, \"Gemini 3.1 Pro Preview\" column"
          },
          {
            "url": "https://thinkingmachines.ai/news/introducing-inkling/",
            "name": "Inkling",
            "title": "Inkling: Our Open-Weights Model",
            "lab": "thinking-machines",
            "date": "2026-07-15",
            "printed": "85.9%",
            "where": "'Benchmarking Inkling' main table, row 'BrowseComp w/ ctx management', column 'Gemini 3.1 Pro (high)'"
          },
          {
            "url": "https://z.ai/blog/glm-5.1",
            "name": "GLM-5.1",
            "title": "GLM-5.1: Towards Long-Horizon Tasks",
            "lab": "zai",
            "date": "2026-04-07",
            "printed": "85.9",
            "where": "Benchmark table, row 'BrowseComp w/ Context Manage', column 'Gemini 3.1 Pro'"
          }
        ],
        "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": "gemini-3-1-pro:swe-bench-pro:default: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": "swe-bench-pro",
        "benchmark_name": "SWE-bench Pro",
        "setting": "default",
        "setting_name": "",
        "score": 54.2,
        "unit": "%",
        "printed": "54.2%",
        "reported_by": "google",
        "run_by": null,
        "label": "Gemini 3.1 Pro",
        "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 'SWE-Bench Pro (Public)', column 'Gemini 3.1 Pro Thinking (High)'",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "single attempt; Thinking (High)",
        "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/claude-fable-5-mythos-5-system-card",
            "name": "Claude Fable 5 and Mythos 5 card",
            "title": "Claude Fable 5 and Claude Mythos 5 System Card",
            "lab": "anthropic",
            "date": "2026-06-09",
            "printed": "54.2",
            "where": "Table 8.1.A (p.251), row SWE-bench Pro, column Gemini 3.1 Pro"
          },
          {
            "url": "https://www.anthropic.com/claude-opus-4-7-system-card",
            "name": "Claude Opus 4.7 card",
            "title": "Claude Opus 4.7 System Card",
            "lab": "anthropic",
            "date": "2026-04-16",
            "printed": "54.2%",
            "where": "Table 8.1.A (p.191), row SWE-bench Pro, column Gemini 3.1 Pro"
          },
          {
            "url": "https://www.anthropic.com/claude-opus-4-8-system-card",
            "name": "Claude Opus 4.8 card",
            "title": "Claude Opus 4.8 System Card",
            "lab": "anthropic",
            "date": "2026-05-28",
            "printed": "54.2",
            "where": "Table 8.1.A (p.194), row SWE-bench Pro, column Gemini 3.1 Pro"
          },
          {
            "url": "https://www.anthropic.com/news/claude-opus-4-7",
            "name": "Claude Opus 4.7",
            "title": "Introducing Claude Opus 4.7",
            "lab": "anthropic",
            "date": "2026-04-16",
            "printed": "54.2%",
            "where": "benchmark table image, row SWE-bench Pro, column Gemini 3.1 Pro"
          },
          {
            "url": "https://www.anthropic.com/news/claude-opus-4-8",
            "name": "Claude Opus 4.8",
            "title": "Introducing Claude Opus 4.8",
            "lab": "anthropic",
            "date": "2026-05-28",
            "printed": "54.2%",
            "where": "benchmark table image, row SWE-Bench Pro, column Gemini 3.1 Pro"
          },
          {
            "url": "https://api-docs.deepseek.com/news/news260424/",
            "name": "DeepSeek V4 Preview",
            "title": "DeepSeek V4 Preview Release",
            "lab": "deepseek",
            "date": "2026-04-24",
            "printed": "54.2",
            "where": "Full benchmark table image; row 'SWE Pro (Resolved)'; column 'Gemini-3.1-Pro High'"
          },
          {
            "url": "https://www.minimax.io/blog/minimax-m3",
            "name": "MiniMax M3",
            "title": "MiniMax M3: Frontier Coding, 1M Context, Native Multimodality — All in One Model",
            "lab": "minimax",
            "date": "2026-06-01",
            "printed": "54.2",
            "where": "Bar chart image (10 panels); panel 'SWE Bench Pro'; value label on Gemini 3.1 Pro bar"
          },
          {
            "url": "https://www.minimax.io/news/minimax-m27-en",
            "name": "MiniMax M2.7",
            "title": "MiniMax M2.7: Early Echoes of Self-Evolution",
            "lab": "minimax",
            "date": "2026-03-18",
            "printed": "54.2",
            "where": "Bar chart image (8 panels); panel 'SWE Bench Pro'; value label on Gemini 3.1 Pro bar"
          },
          {
            "url": "https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-5/",
            "name": "Gemini 3.5 Flash",
            "title": "Gemini 3.5: frontier intelligence with action",
            "lab": "google",
            "date": "2026-05-19",
            "printed": "54.2%",
            "where": "Benchmark table image; row 'SWE-Bench Pro (Public), Single attempt'; column gemini-3-1-pro"
          },
          {
            "url": "https://deepmind.google/models/model-cards/gemini-3-6-flash/",
            "name": "Gemini 3.6 Flash and 3.5 Flash-Lite card",
            "title": "Gemini 3.6 Flash - Model Card",
            "lab": "google",
            "date": "2026-07-21",
            "printed": "54.2%",
            "where": "HTML results table 'Results as of July, 2026'; row 'SWE-Bench Pro (Public)'; column gemini-3-1-pro"
          },
          {
            "url": "https://www.kimi.ai/blog/kimi-k2-6",
            "name": "Kimi K2.6",
            "title": "Kimi K2.6: Advancing Open-Source Coding",
            "lab": "moonshot",
            "date": "2026-04-20",
            "printed": "54.2",
            "where": "Benchmark Table (appendix), row SWE-Bench Pro ; column Gemini 3.1 Pro (thinking high)"
          },
          {
            "url": "https://openai.com/index/introducing-gpt-5-5",
            "name": "GPT-5.5",
            "title": "Introducing GPT-5.5",
            "lab": "openai",
            "date": "2026-04-23",
            "printed": "54.2%",
            "where": "table 1 \"SWE-Bench Pro (Public) *\" row, \"Gemini 3.1 Pro\" column"
          },
          {
            "url": "https://openai.com/index/gpt-5-6",
            "name": "GPT-5.6",
            "title": "GPT-5.6: Frontier intelligence that scales with your ambition",
            "lab": "openai",
            "date": "2026-07-09",
            "printed": "54.2%",
            "where": "table 1 \"SWE-Bench Pro\" row, \"Gemini 3.1 Pro Preview\" column"
          },
          {
            "url": "https://thinkingmachines.ai/news/introducing-inkling/",
            "name": "Inkling",
            "title": "Inkling: Our Open-Weights Model",
            "lab": "thinking-machines",
            "date": "2026-07-15",
            "printed": "54.2%",
            "where": "'Benchmarking Inkling' main table, row 'SWEBench Pro Public', column 'Gemini 3.1 Pro (high)'"
          },
          {
            "url": "https://z.ai/blog/glm-5.1",
            "name": "GLM-5.1",
            "title": "GLM-5.1: Towards Long-Horizon Tasks",
            "lab": "zai",
            "date": "2026-04-07",
            "printed": "54.2",
            "where": "Tabbed bar chart at top, tab 'Software Engineering' (SWE-Bench Pro), bar 'Gemini 3.1 Pro'"
          },
          {
            "url": "https://z.ai/blog/glm-5.2",
            "name": "GLM-5.2",
            "title": "GLM-5.2: Built for Long-Horizon Tasks",
            "lab": "zai",
            "date": "2026-06-16",
            "printed": "54.2",
            "where": "Full Benchmark Table, row 'SWE-bench Pro', column 'Gemini 3.1 Pro'"
          }
        ],
        "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": "gemini-3-1-pro:deepswe-1-1:default: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": "deepswe-1-1",
        "benchmark_name": "DeepSWE v1.1",
        "setting": "default",
        "setting_name": "",
        "score": 12.0,
        "unit": "%",
        "printed": "12%",
        "reported_by": "google",
        "run_by": null,
        "label": "3.1 Pro",
        "source": "https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-6-flash-3-5-flash-lite-3-5-flash-cyber/",
        "source_title": "Introducing Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber",
        "source_date": "2026-07-21",
        "where": "Gemini 3.6 Flash post, chart 'Outperforms prior generations across agentic benchmarks' (image), panel 'DeepSWE v1.1', bar '3.1 Pro' value label",
        "own": false,
        "primary": true,
        "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://deepmind.google/models/model-cards/gemini-3-6-flash/",
            "name": "Gemini 3.6 Flash and 3.5 Flash-Lite card",
            "title": "Gemini 3.6 Flash - Model Card",
            "lab": "google",
            "date": "2026-07-21",
            "printed": "12%",
            "where": "HTML results table 'Results as of July, 2026'; row 'DeepSWE v1.1'; column gemini-3-1-pro"
          }
        ],
        "source_image": "https://frontierlog.ai/sources/eed20d4f60c90b41.webp",
        "source_image_original": "https://storage.googleapis.com/gweb-uniblog-publish-prod/images/gemini-3-6-flash__evals__figure-.width-2000.format-webp.webp"
      }
    ]
  }
}