{
  "api_version": "1",
  "updated": "2026-10-26",
  "latest_event": "2026-10-24",
  "rights": "© Frontier Log, all rights reserved.",
  "model": {
    "slug": "claude-opus-4-5",
    "name": "Claude Opus 4.5",
    "lab": "anthropic",
    "date": "2025-11-24",
    "source": "https://www.anthropic.com/news/claude-opus-4-5",
    "source_title": "Introducing Claude Opus 4.5",
    "flagship": true,
    "event": "https://frontierlog.ai/event/2025-11-24-claude-opus-4-5",
    "releases": [],
    "variant_of": null,
    "made_by": "Anthropic",
    "url": "https://frontierlog.ai/model/claude-opus-4-5",
    "results": [
      {
        "id": "claude-opus-4-5:swe-bench-verified:default: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": "swe-bench-verified",
        "benchmark_name": "SWE-bench Verified",
        "setting": "default",
        "setting_name": "",
        "score": 80.9,
        "unit": "%",
        "printed": "80.9%",
        "reported_by": "anthropic",
        "run_by": null,
        "label": "Opus 4.5",
        "source": "https://www.anthropic.com/news/claude-opus-4-5",
        "source_title": "Introducing Claude Opus 4.5",
        "source_date": "2025-11-24",
        "where": "benchmark table (image), row 'Agentic coding, SWE-bench Verified', column 'Opus 4.5'",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "no thinking budget, high effort; averaged over 5 trials",
        "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-4-5-system-card",
            "name": "Claude Opus 4.5 card",
            "title": "Claude Opus 4.5 System Card",
            "lab": "anthropic",
            "date": "2025-11-24",
            "printed": "80.9%",
            "where": "Table 2.3.A (p.19), row SWE-bench Verified, column Claude Opus 4.5 (footnote: without extended thinking)"
          },
          {
            "url": "https://www.anthropic.com/claude-opus-4-6-system-card",
            "name": "Claude Opus 4.6 card",
            "title": "Claude Opus 4.6 System Card",
            "lab": "anthropic",
            "date": "2026-02-05",
            "printed": "80.9%",
            "where": "Table 2.3.A (p.18), row SWE-bench Verified, column Claude Opus 4.5"
          },
          {
            "url": "https://www.anthropic.com/news/claude-opus-4-6",
            "name": "Claude Opus 4.6",
            "title": "Claude Opus 4.6",
            "lab": "anthropic",
            "date": "2026-02-05",
            "printed": "80.9%",
            "where": "Benchmark table image, row 'Agentic coding SWE-bench Verified', column 'Opus 4.5'"
          },
          {
            "url": "https://www.minimax.io/news/minimax-m21",
            "name": "MiniMax M2.1",
            "title": "MiniMax M2.1: Significantly Enhanced Multi-Language Programming, Built for Real-World Complex Tasks",
            "lab": "minimax",
            "date": "2025-12-23",
            "printed": "80.9",
            "where": "Coding table image; row 'SWE-bench Verified'; column 'Claude Opus 4.5'"
          },
          {
            "url": "https://www.minimax.io/news/minimax-m25",
            "name": "MiniMax M2.5",
            "title": "MiniMax M2.5: Built for Real-World Productivity.",
            "lab": "minimax",
            "date": "2026-02-12",
            "printed": "80.9",
            "where": "Overview bar chart image (8 panels); panel 'SWE-Bench Verified'; value label on Claude Opus 4.5 bar"
          },
          {
            "url": "https://www.kimi.ai/blog/kimi-k2-5",
            "name": "Kimi K2.5",
            "title": "Kimi K2.5: Visual Agentic Intelligence",
            "lab": "moonshot",
            "date": "2026-01-27",
            "printed": "80.9",
            "where": "Appendix Benchmark table, row SWE-Bench Verified ; column Claude 4.5 Opus (Extend Thinking)"
          },
          {
            "url": "https://qwen.ai/blog?id=qwen3-max-thinking",
            "name": "Qwen3-Max-Thinking",
            "title": "Pushing Qwen3-Max-Thinking Beyond its Limits",
            "lab": "alibaba",
            "date": "2026-01-26",
            "printed": "80.9",
            "where": "Score chart image (score.png), panel SWE-bench Verified, value label on the Claude-Opus-4.5 bar"
          },
          {
            "url": "https://z.ai/blog/glm-5",
            "name": "GLM-5",
            "title": "GLM-5: From Vibe Coding to Agentic Engineering",
            "lab": "zai",
            "date": "2026-02-12",
            "printed": "80.9",
            "where": "Benchmark table, row 'SWE-bench Verified', column 'Claude Opus 4.5 (Extend Thinking)'"
          }
        ],
        "source_image": "https://frontierlog.ai/sources/790b0c5480216588.png",
        "source_image_original": "https://www-cdn.anthropic.com/images/4zrzovbb/website/52303b11db76017fd0c2f73c7fafa5c752515979-2600x2236.png"
      },
      {
        "id": "claude-opus-4-5:terminal-bench-2-0:default: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": "terminal-bench-2-0",
        "benchmark_name": "Terminal-Bench 2.0",
        "setting": "default",
        "setting_name": "",
        "score": 59.3,
        "unit": "%",
        "printed": "59.3%",
        "reported_by": "anthropic",
        "run_by": null,
        "label": "Opus 4.5",
        "source": "https://www.anthropic.com/news/claude-opus-4-5",
        "source_title": "Introducing Claude Opus 4.5",
        "source_date": "2025-11-24",
        "where": "benchmark table (image), row 'Agentic terminal coding, Terminal-bench 2.0', column 'Opus 4.5'",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "128K thinking budget, high effort; averaged over 5 trials",
        "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-4-5-system-card",
            "name": "Claude Opus 4.5 card",
            "title": "Claude Opus 4.5 System Card",
            "lab": "anthropic",
            "date": "2025-11-24",
            "printed": "59.3%",
            "where": "Table 2.3.A (p.19), row Terminal-bench 2.0, column Claude Opus 4.5 (footnote: 128k thinking budget)"
          },
          {
            "url": "https://www.kimi.ai/blog/kimi-k2-5",
            "name": "Kimi K2.5",
            "title": "Kimi K2.5: Visual Agentic Intelligence",
            "lab": "moonshot",
            "date": "2026-01-27",
            "printed": "59.3",
            "where": "Appendix Benchmark table, row Terminal-Bench 2.0 (Terminus-2; K2.5 non-thinking mode) ; column Claude 4.5 Opus (Extend Thinking)"
          },
          {
            "url": "https://qwen.ai/blog?id=qwen3.5",
            "name": "Qwen3.5",
            "title": "Qwen3.5: Towards Native Multimodal Agents",
            "lab": "alibaba",
            "date": "2026-02-16",
            "printed": "59.3",
            "where": "Score chart image (qwen3.5_397b_a17b_score.png), panel Terminal-Bench 2, value label on the Claude Opus 4.5 bar"
          },
          {
            "url": "https://z.ai/blog/glm-5",
            "name": "GLM-5",
            "title": "GLM-5: From Vibe Coding to Agentic Engineering",
            "lab": "zai",
            "date": "2026-02-12",
            "printed": "59.3",
            "where": "Benchmark table, row 'Terminal-Bench 2.0 Terminus-2 (GLM-5 cell '56.2 / 60.7†': first value; 60.7 is the verified variant)', column 'Claude Opus 4.5 (Extend Thinking)'"
          }
        ],
        "source_image": "https://frontierlog.ai/sources/790b0c5480216588.png",
        "source_image_original": "https://www-cdn.anthropic.com/images/4zrzovbb/website/52303b11db76017fd0c2f73c7fafa5c752515979-2600x2236.png"
      },
      {
        "id": "claude-opus-4-5:arc-agi-2:default: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": "arc-agi-2",
        "benchmark_name": "ARC-AGI-2",
        "setting": "default",
        "setting_name": "",
        "score": 37.6,
        "unit": "%",
        "printed": "37.6%",
        "reported_by": "anthropic",
        "run_by": null,
        "label": "Opus 4.5",
        "source": "https://www.anthropic.com/news/claude-opus-4-5",
        "source_title": "Introducing Claude Opus 4.5",
        "source_date": "2025-11-24",
        "where": "benchmark table (image), row 'Novel problem solving, ARC-AGI-2 (Verified)', column 'Opus 4.5'",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "64K thinking budget, high effort; averaged over 5 trials",
        "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-4-5-system-card",
            "name": "Claude Opus 4.5 card",
            "title": "Claude Opus 4.5 System Card",
            "lab": "anthropic",
            "date": "2025-11-24",
            "printed": "37.6%",
            "where": "Table 2.3.A (p.19), row ARC-AGI-2 (Verified), column Claude Opus 4.5"
          },
          {
            "url": "https://www.anthropic.com/claude-opus-4-6-system-card",
            "name": "Claude Opus 4.6 card",
            "title": "Claude Opus 4.6 System Card",
            "lab": "anthropic",
            "date": "2026-02-05",
            "printed": "37.6%",
            "where": "Table 2.3.A (p.18), row ARC-AGI-2 (Verified), column Claude Opus 4.5"
          },
          {
            "url": "https://www.anthropic.com/news/claude-opus-4-6",
            "name": "Claude Opus 4.6",
            "title": "Claude Opus 4.6",
            "lab": "anthropic",
            "date": "2026-02-05",
            "printed": "37.6%",
            "where": "Benchmark table image, row 'Novel problem-solving ARC AGI 2', column 'Opus 4.5'"
          }
        ],
        "source_image": "https://frontierlog.ai/sources/790b0c5480216588.png",
        "source_image_original": "https://www-cdn.anthropic.com/images/4zrzovbb/website/52303b11db76017fd0c2f73c7fafa5c752515979-2600x2236.png"
      },
      {
        "id": "claude-opus-4-5:gpqa-diamond:default: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": "gpqa-diamond",
        "benchmark_name": "GPQA Diamond",
        "setting": "default",
        "setting_name": "",
        "score": 87.0,
        "unit": "%",
        "printed": "87.0%",
        "reported_by": "anthropic",
        "run_by": null,
        "label": "Opus 4.5",
        "source": "https://www.anthropic.com/news/claude-opus-4-5",
        "source_title": "Introducing Claude Opus 4.5",
        "source_date": "2025-11-24",
        "where": "benchmark table (image), row 'Graduate-level reasoning, GPQA Diamond', column 'Opus 4.5'",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "64K thinking budget, high effort; averaged over 5 trials",
        "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-4-5-system-card",
            "name": "Claude Opus 4.5 card",
            "title": "Claude Opus 4.5 System Card",
            "lab": "anthropic",
            "date": "2025-11-24",
            "printed": "87.0%",
            "where": "Table 2.3.A (p.19), row GPQA Diamond, column Claude Opus 4.5"
          },
          {
            "url": "https://www.anthropic.com/claude-opus-4-6-system-card",
            "name": "Claude Opus 4.6 card",
            "title": "Claude Opus 4.6 System Card",
            "lab": "anthropic",
            "date": "2026-02-05",
            "printed": "87.0%",
            "where": "Table 2.3.A (p.18), row GPQA Diamond, column Claude Opus 4.5"
          },
          {
            "url": "https://www.anthropic.com/news/claude-opus-4-6",
            "name": "Claude Opus 4.6",
            "title": "Claude Opus 4.6",
            "lab": "anthropic",
            "date": "2026-02-05",
            "printed": "87.0%",
            "where": "Benchmark table image, row 'Graduate-level reasoning GPQA Diamond', column 'Opus 4.5'"
          },
          {
            "url": "https://www.minimax.io/news/minimax-m21",
            "name": "MiniMax M2.1",
            "title": "MiniMax M2.1: Significantly Enhanced Multi-Language Programming, Built for Real-World Complex Tasks",
            "lab": "minimax",
            "date": "2025-12-23",
            "printed": "87.0",
            "where": "Tool-use and intelligence table image; row 'GPQA-D'; column 'Claude Opus 4.5'"
          },
          {
            "url": "https://www.minimax.io/news/minimax-m25",
            "name": "MiniMax M2.5",
            "title": "MiniMax M2.5: Built for Real-World Productivity.",
            "lab": "minimax",
            "date": "2026-02-12",
            "printed": "87.0",
            "where": "Appendix table image; row 'GPQA-D'; column 'Claude Opus 4.5'"
          },
          {
            "url": "https://www.kimi.ai/blog/kimi-k2-5",
            "name": "Kimi K2.5",
            "title": "Kimi K2.5: Visual Agentic Intelligence",
            "lab": "moonshot",
            "date": "2026-01-27",
            "printed": "87.0",
            "where": "Appendix Benchmark table, row GPQA-Diamond (avg@8) ; column Claude 4.5 Opus (Extend Thinking)"
          },
          {
            "url": "https://qwen.ai/blog?id=qwen3-max-thinking",
            "name": "Qwen3-Max-Thinking",
            "title": "Pushing Qwen3-Max-Thinking Beyond its Limits",
            "lab": "alibaba",
            "date": "2026-01-26",
            "printed": "87.0",
            "where": "Score chart image (score.png), panel GPQA Diamond, value label on the Claude-Opus-4.5 bar"
          },
          {
            "url": "https://z.ai/blog/glm-5",
            "name": "GLM-5",
            "title": "GLM-5: From Vibe Coding to Agentic Engineering",
            "lab": "zai",
            "date": "2026-02-12",
            "printed": "87.0",
            "where": "Benchmark table, row 'GPQA-Diamond', column 'Claude Opus 4.5 (Extend Thinking)'"
          }
        ],
        "source_image": "https://frontierlog.ai/sources/790b0c5480216588.png",
        "source_image_original": "https://www-cdn.anthropic.com/images/4zrzovbb/website/52303b11db76017fd0c2f73c7fafa5c752515979-2600x2236.png"
      },
      {
        "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",
        "benchmark_name": "Humanity's Last Exam",
        "setting": "no-tools",
        "setting_name": "No tools",
        "score": 30.8,
        "unit": "%",
        "printed": "30.8%",
        "reported_by": "anthropic",
        "run_by": null,
        "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; Claude Sonnet 4.5 as grader",
        "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/news/claude-opus-4-6",
            "name": "Claude Opus 4.6",
            "title": "Claude Opus 4.6",
            "lab": "anthropic",
            "date": "2026-02-05",
            "printed": "30.8%",
            "where": "Benchmark table image, row 'Multidisciplinary reasoning Humanity's Last Exam, 'without tools'', column 'Opus 4.5'"
          },
          {
            "url": "https://qwen.ai/blog?id=qwen3-max-thinking",
            "name": "Qwen3-Max-Thinking",
            "title": "Pushing Qwen3-Max-Thinking Beyond its Limits",
            "lab": "alibaba",
            "date": "2026-01-26",
            "printed": "30.8",
            "where": "Score chart image (score.png), panel Humanity's Last Exam, value label on the Claude-Opus-4.5 bar"
          }
        ],
        "source_capture": "https://frontierlog.ai/captures/119c3ba76f816d86.png"
      },
      {
        "id": "claude-opus-4-5:hle:with-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",
        "benchmark_name": "Humanity's Last Exam",
        "setting": "with-tools",
        "setting_name": "With tools",
        "score": 43.2,
        "unit": "%",
        "printed": "43.2%",
        "reported_by": "anthropic",
        "run_by": null,
        "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 'With Search', bar 'Opus 4.5' (page 32)",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "web search, web fetch and code execution, no reasoning; Claude Sonnet 4.5 as grader",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-17",
        "check": "human",
        "corroborated": true,
        "corroborated_on": "2026-10-19",
        "corroborated_by": [
          {
            "url": "https://www.kimi.ai/blog/kimi-k2-5",
            "name": "Kimi K2.5",
            "title": "Kimi K2.5: Visual Agentic Intelligence",
            "lab": "moonshot",
            "date": "2026-01-27",
            "printed": "43.2",
            "where": "Appendix Benchmark table, row HLE-Full w/ tools ; column Claude 4.5 Opus (Extend Thinking)"
          },
          {
            "url": "https://qwen.ai/blog?id=qwen3-max-thinking",
            "name": "Qwen3-Max-Thinking",
            "title": "Pushing Qwen3-Max-Thinking Beyond its Limits",
            "lab": "alibaba",
            "date": "2026-01-26",
            "printed": "43.2",
            "where": "Score chart image (score.png), panel Humanity's Last Exam (with Search), value label on the Claude-Opus-4.5 bar"
          }
        ],
        "source_capture": "https://frontierlog.ai/captures/119c3ba76f816d86.png"
      },
      {
        "id": "claude-opus-4-5:hle:with-tools:2",
        "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",
        "benchmark_name": "Humanity's Last Exam",
        "setting": "with-tools",
        "setting_name": "With tools",
        "score": 43.4,
        "unit": "%",
        "printed": "43.4%",
        "reported_by": "anthropic",
        "run_by": null,
        "label": "Opus 4.5",
        "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 4.5', 'with tools'",
        "own": false,
        "primary": false,
        "parallel": false,
        "notes": "web search, web fetch, code execution, programmatic tool calling; compaction at 50K up to 3M tokens; max effort, adaptive thinking",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-17",
        "check": "human",
        "corroborated": true,
        "corroborated_on": "2026-10-19",
        "corroborated_by": [
          {
            "url": "https://z.ai/blog/glm-5",
            "name": "GLM-5",
            "title": "GLM-5: From Vibe Coding to Agentic Engineering",
            "lab": "zai",
            "date": "2026-02-12",
            "printed": "43.4",
            "where": "Benchmark table, row 'Humanity's Last Exam w/ Tools (* = full set for Opus 4.5, Gemini, GPT-5.2)', column 'Claude Opus 4.5 (Extend Thinking)'"
          }
        ],
        "source_image": "https://frontierlog.ai/sources/e6c54904fe1e4f51.png",
        "source_image_original": "https://www-cdn.anthropic.com/images/4zrzovbb/website/f9564dd2f758237bd9dbe775674c4a375aff1e8a-2600x2968.png"
      },
      {
        "id": "claude-opus-4-5:browsecomp:default: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": "browsecomp",
        "benchmark_name": "BrowseComp",
        "setting": "default",
        "setting_name": "",
        "score": 67.8,
        "unit": "%",
        "printed": "67.8%",
        "reported_by": "anthropic",
        "run_by": null,
        "label": "Opus 4.5",
        "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 'Agentic search, BrowseComp', column 'Opus 4.5'",
        "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://www.minimax.io/news/minimax-m25",
            "name": "MiniMax M2.5",
            "title": "MiniMax M2.5: Built for Real-World Productivity.",
            "lab": "minimax",
            "date": "2026-02-12",
            "printed": "67.8",
            "where": "Overview bar chart image (8 panels); panel 'BrowseComp (w/ctx)'; value label on Claude Opus 4.5 bar"
          },
          {
            "url": "https://qwen.ai/blog?id=qwen3.5",
            "name": "Qwen3.5",
            "title": "Qwen3.5: Towards Native Multimodal Agents",
            "lab": "alibaba",
            "date": "2026-02-16",
            "printed": "67.8",
            "where": "Score chart image (qwen3.5_397b_a17b_score.png), panel BrowseComp, value label on the Claude Opus 4.5 bar"
          },
          {
            "url": "https://z.ai/blog/glm-5",
            "name": "GLM-5",
            "title": "GLM-5: From Vibe Coding to Agentic Engineering",
            "lab": "zai",
            "date": "2026-02-12",
            "printed": "67.8",
            "where": "Benchmark table, row 'BrowseComp w/ Context Manage', column 'Claude Opus 4.5 (Extend Thinking)'"
          }
        ],
        "source_image": "https://frontierlog.ai/sources/e6c54904fe1e4f51.png",
        "source_image_original": "https://www-cdn.anthropic.com/images/4zrzovbb/website/f9564dd2f758237bd9dbe775674c4a375aff1e8a-2600x2968.png"
      },
      {
        "id": "claude-opus-4-5:swe-bench-pro:default: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": "swe-bench-pro",
        "benchmark_name": "SWE-bench Pro",
        "setting": "default",
        "setting_name": "",
        "score": 52.0,
        "unit": "%",
        "printed": "52.0%",
        "reported_by": "anthropic",
        "run_by": null,
        "label": "Claude Opus 4.5 (no thinking)",
        "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": "Section 2.4, Table 2.4.A, column 'SWE-bench Pro', row 'Claude Opus 4.5 (no thinking)'",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "no thinking; average of 5 trials",
        "caveat": "The source doesn't say whether this is the public or the commercial set",
        "human_verified": true,
        "human_verified_on": "2026-10-19",
        "check": "human",
        "corroborated": false,
        "corroborated_on": null,
        "corroborated_by": [],
        "source_capture": "https://frontierlog.ai/captures/ceb9c141f42aab5e.png"
      },
      {
        "id": "claude-opus-4-5:swe-bench-pro:default:2",
        "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": "swe-bench-pro",
        "benchmark_name": "SWE-bench Pro",
        "setting": "default",
        "setting_name": "",
        "score": 51.6,
        "unit": "%",
        "printed": "51.60%",
        "reported_by": "anthropic",
        "run_by": null,
        "label": "Claude Opus 4.5 (64k thinking)",
        "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": "Section 2.4, Table 2.4.A, column 'SWE-bench Pro', row 'Claude Opus 4.5 (64k thinking)'",
        "own": true,
        "primary": false,
        "parallel": false,
        "notes": "64K thinking budget; average of 5 trials",
        "caveat": "The source doesn't say whether this is the public or the commercial set",
        "human_verified": true,
        "human_verified_on": "2026-10-19",
        "check": "human",
        "corroborated": false,
        "corroborated_on": null,
        "corroborated_by": [],
        "source_capture": "https://frontierlog.ai/captures/ceb9c141f42aab5e.png"
      },
      {
        "id": "claude-opus-4-5:terminal-bench-2-0:default:2",
        "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": "terminal-bench-2-0",
        "benchmark_name": "Terminal-Bench 2.0",
        "setting": "default",
        "setting_name": "",
        "score": 59.8,
        "unit": "%",
        "printed": "59.8%",
        "reported_by": "anthropic",
        "run_by": null,
        "label": "Claude Opus 4.5",
        "source": "https://www.anthropic.com/claude-opus-4-6-system-card",
        "source_title": "Claude Opus 4.6 System Card",
        "source_date": "2026-02-05",
        "where": "Table 2.3.A (PDF page 18), row 'Terminal-Bench 2.0', column 'Claude Opus 4.5'",
        "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://www.anthropic.com/news/claude-opus-4-6",
            "name": "Claude Opus 4.6",
            "title": "Claude Opus 4.6",
            "lab": "anthropic",
            "date": "2026-02-05",
            "printed": "59.8%",
            "where": "main benchmark table image, row 'Agentic terminal coding / Terminal-Bench 2.0', column 'Opus 4.5'"
          },
          {
            "url": "https://www.anthropic.com/news/claude-sonnet-4-6",
            "name": "Claude Sonnet 4.6",
            "title": "Introducing Sonnet 4.6",
            "lab": "anthropic",
            "date": "2026-02-17",
            "printed": "59.8%",
            "where": "main benchmark table image, row 'Agentic terminal coding / Terminal-Bench 2.0', column 'Opus 4.5'"
          },
          {
            "url": "https://www.anthropic.com/claude-sonnet-4-6-system-card",
            "name": "Claude Sonnet 4.6 card",
            "title": "Claude Sonnet 4.6 System Card",
            "lab": "anthropic",
            "date": "2026-02-17",
            "printed": "59.8%",
            "where": "section 2 evaluation summary table, row 'Terminal-Bench 2.0', column 'Claude Opus 4.5'"
          }
        ],
        "source_capture": "https://frontierlog.ai/captures/32cbf86df2814c55.png"
      }
    ]
  }
}