{
  "api_version": "1",
  "updated": "2026-10-26",
  "latest_event": "2026-10-24",
  "rights": "© Frontier Log, all rights reserved.",
  "model": {
    "slug": "claude-opus-4",
    "name": "Claude Opus 4",
    "lab": "anthropic",
    "date": "2025-05-22",
    "flagship": true,
    "event": "https://frontierlog.ai/event/2025-05-22-claude-4",
    "releases": [],
    "variant_of": null,
    "made_by": "Anthropic",
    "url": "https://frontierlog.ai/model/claude-opus-4",
    "results": [
      {
        "id": "claude-opus-4:swe-bench-verified:default:1",
        "model": "claude-opus-4",
        "model_name": "Claude Opus 4",
        "model_date": "2025-05-22",
        "event": "https://frontierlog.ai/event/2025-05-22-claude-4",
        "benchmark": "swe-bench-verified",
        "benchmark_name": "SWE-bench Verified",
        "setting": "default",
        "setting_name": "",
        "score": 72.5,
        "unit": "%",
        "printed": "72.5%",
        "reported_by": "anthropic",
        "run_by": null,
        "label": "Opus 4",
        "source": "https://www.anthropic.com/news/claude-4",
        "source_title": "Introducing Claude 4",
        "source_date": "2025-05-22",
        "where": "'Software engineering, SWE-bench verified' bar chart (image), bar 'Opus 4', value inside the bar",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "no extended thinking; bash and file-editing tools; full 500 problems",
        "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": "72.5%",
            "where": "Benchmark table image, row 'Agentic coding SWE-bench Verified', column 'Claude Opus 4'"
          },
          {
            "url": "https://www.minimax.io/news/minimaxm1",
            "name": "MiniMax-M1",
            "title": "MiniMax-M1, the World's First Open-Source, Large-Scale, Hybrid-Attention Reasoning Model",
            "lab": "minimax",
            "date": "2025-06-16",
            "printed": "72.5",
            "where": "Full benchmark table image; row 'SWE-bench Verified'; column 'Claude 4 Opus (64K thinking)'"
          },
          {
            "url": "https://www.kimi.ai/blog/kimi-k2",
            "name": "Kimi K2",
            "title": "Kimi K2: Open Agentic Intelligence",
            "lab": "moonshot",
            "date": "2025-07-11",
            "printed": "72.5",
            "where": "Evaluation Results table (Kimi-K2-Instruct tab), row SWE-bench Verified (Agentic Coding) Single Attempt (Acc) ; column Claude Opus 4 (w/o extended thinking)"
          },
          {
            "url": "https://qwen.ai/blog?id=qwen3-max",
            "name": "Qwen3-Max",
            "title": "Qwen3-Max: Just Scale it",
            "lab": "alibaba",
            "date": "2025-09-24",
            "printed": "72.5",
            "where": "Qwen3-Max-Instruct bar chart (Qwen3-Max-Instruct.png), group SWE-Bench (Verified), value label on the Claude Opus 4 Non-thinking bar"
          }
        ],
        "source_image": "https://frontierlog.ai/sources/5ed59708a2aeef89.png",
        "source_image_original": "https://www-cdn.anthropic.com/images/4zrzovbb/website/09a6d5aa47c25cb2037efff9f486da4918f77708-3840x2304.png"
      },
      {
        "id": "claude-opus-4:swe-bench-verified:default:2",
        "model": "claude-opus-4",
        "model_name": "Claude Opus 4",
        "model_date": "2025-05-22",
        "event": "https://frontierlog.ai/event/2025-05-22-claude-4",
        "benchmark": "swe-bench-verified",
        "benchmark_name": "SWE-bench Verified",
        "setting": "default",
        "setting_name": "",
        "score": 79.4,
        "unit": "%",
        "printed": "79.4%",
        "reported_by": "anthropic",
        "run_by": null,
        "label": "Opus 4",
        "source": "https://www.anthropic.com/news/claude-4",
        "source_title": "Introducing Claude 4",
        "source_date": "2025-05-22",
        "where": "'Software engineering, SWE-bench verified' bar chart (image), bar 'Opus 4', label above the bar",
        "own": true,
        "primary": false,
        "parallel": true,
        "notes": "parallel test-time compute",
        "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",
            "name": "Kimi K2",
            "title": "Kimi K2: Open Agentic Intelligence",
            "lab": "moonshot",
            "date": "2025-07-11",
            "printed": "79.4",
            "where": "Evaluation Results table (Kimi-K2-Instruct tab), row SWE-bench Verified (Agentic Coding) Multiple Attempts (Acc) ; column Claude Opus 4 (w/o extended thinking)"
          }
        ],
        "source_image": "https://frontierlog.ai/sources/5ed59708a2aeef89.png",
        "source_image_original": "https://www-cdn.anthropic.com/images/4zrzovbb/website/09a6d5aa47c25cb2037efff9f486da4918f77708-3840x2304.png"
      },
      {
        "id": "claude-opus-4:gpqa-diamond:default:1",
        "model": "claude-opus-4",
        "model_name": "Claude Opus 4",
        "model_date": "2025-05-22",
        "event": "https://frontierlog.ai/event/2025-05-22-claude-4",
        "benchmark": "gpqa-diamond",
        "benchmark_name": "GPQA Diamond",
        "setting": "default",
        "setting_name": "",
        "score": 74.9,
        "unit": "%",
        "printed": "74.9%",
        "reported_by": "anthropic",
        "run_by": null,
        "label": "Opus 4",
        "source": "https://www.anthropic.com/news/claude-4",
        "source_title": "Introducing Claude 4",
        "source_date": "2025-05-22",
        "where": "Appendix, 'Performance benchmark reporting', extended-thinking list, GPQA Diamond item",
        "own": true,
        "primary": false,
        "parallel": false,
        "notes": "no extended thinking",
        "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",
            "name": "Kimi K2",
            "title": "Kimi K2: Open Agentic Intelligence",
            "lab": "moonshot",
            "date": "2025-07-11",
            "printed": "74.9",
            "where": "Evaluation Results table (Kimi-K2-Instruct tab), row GPQA-Diamond Avg@8 ; column Claude Opus 4 (w/o extended thinking)"
          }
        ]
      },
      {
        "id": "claude-opus-4:gpqa-diamond:default:2",
        "model": "claude-opus-4",
        "model_name": "Claude Opus 4",
        "model_date": "2025-05-22",
        "event": "https://frontierlog.ai/event/2025-05-22-claude-4",
        "benchmark": "gpqa-diamond",
        "benchmark_name": "GPQA Diamond",
        "setting": "default",
        "setting_name": "",
        "score": 79.6,
        "unit": "%",
        "printed": "79.6%",
        "reported_by": "anthropic",
        "run_by": null,
        "label": "Claude Opus 4",
        "source": "https://www.anthropic.com/news/claude-4",
        "source_title": "Introducing Claude 4",
        "source_date": "2025-05-22",
        "where": "benchmark table drawn by the page's script (not in the HTML source; render the page), row 'Graduate-level reasoning, GPQA Diamond', column 'Claude Opus 4', first value of the cell",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "extended thinking up to 64K tokens",
        "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": "79.6%",
            "where": "Benchmark table image, row 'Graduate-level reasoning GPQA Diamond', column 'Claude Opus 4'"
          },
          {
            "url": "https://www.minimax.io/news/minimaxm1",
            "name": "MiniMax-M1",
            "title": "MiniMax-M1, the World's First Open-Source, Large-Scale, Hybrid-Attention Reasoning Model",
            "lab": "minimax",
            "date": "2025-06-16",
            "printed": "79.6",
            "where": "Full benchmark table image; row 'GPQA Diamond'; column 'Claude 4 Opus (64K thinking)'"
          }
        ],
        "source_capture": "https://frontierlog.ai/captures/4e6f6af599737101.png"
      },
      {
        "id": "claude-opus-4:gpqa-diamond:default:3",
        "model": "claude-opus-4",
        "model_name": "Claude Opus 4",
        "model_date": "2025-05-22",
        "event": "https://frontierlog.ai/event/2025-05-22-claude-4",
        "benchmark": "gpqa-diamond",
        "benchmark_name": "GPQA Diamond",
        "setting": "default",
        "setting_name": "",
        "score": 83.3,
        "unit": "%",
        "printed": "83.3%",
        "reported_by": "anthropic",
        "run_by": null,
        "label": "Claude Opus 4",
        "source": "https://www.anthropic.com/news/claude-4",
        "source_title": "Introducing Claude 4",
        "source_date": "2025-05-22",
        "where": "benchmark table drawn by the page's script (not in the HTML source; render the page), row 'Graduate-level reasoning, GPQA Diamond', column 'Claude Opus 4', second value of the cell",
        "own": true,
        "primary": false,
        "parallel": true,
        "notes": "parallel test-time compute: best of several samples by an internal scoring model",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-17",
        "check": "human",
        "corroborated": false,
        "corroborated_on": null,
        "corroborated_by": [],
        "source_capture": "https://frontierlog.ai/captures/4e6f6af599737101.png"
      },
      {
        "id": "claude-opus-4:terminal-bench-1-0:default:1",
        "model": "claude-opus-4",
        "model_name": "Claude Opus 4",
        "model_date": "2025-05-22",
        "event": "https://frontierlog.ai/event/2025-05-22-claude-4",
        "benchmark": "terminal-bench-1-0",
        "benchmark_name": "Terminal-Bench 1.0",
        "setting": "default",
        "setting_name": "",
        "score": 43.2,
        "unit": "%",
        "printed": "43.2%",
        "reported_by": "anthropic",
        "run_by": null,
        "label": "Claude Opus 4",
        "source": "https://www.anthropic.com/news/claude-4",
        "source_title": "Introducing Claude 4",
        "source_date": "2025-05-22",
        "where": "main benchmark table (HTML drawn by script), row 'Agentic terminal coding / Terminal-bench', column 'Claude Opus 4', first number of the cell",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "no extended thinking; Claude Code harness",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-18",
        "check": "human",
        "corroborated": false,
        "corroborated_on": null,
        "corroborated_by": [],
        "source_capture": "https://frontierlog.ai/captures/4e6f6af599737101.png"
      },
      {
        "id": "claude-opus-4:terminal-bench-1-0:default:2",
        "model": "claude-opus-4",
        "model_name": "Claude Opus 4",
        "model_date": "2025-05-22",
        "event": "https://frontierlog.ai/event/2025-05-22-claude-4",
        "benchmark": "terminal-bench-1-0",
        "benchmark_name": "Terminal-Bench 1.0",
        "setting": "default",
        "setting_name": "",
        "score": 50.0,
        "unit": "%",
        "printed": "50.0%",
        "reported_by": "anthropic",
        "run_by": null,
        "label": "Claude Opus 4",
        "source": "https://www.anthropic.com/news/claude-4",
        "source_title": "Introducing Claude 4",
        "source_date": "2025-05-22",
        "where": "main benchmark table, row 'Terminal-bench', column 'Claude Opus 4', second number of the cell (after the slash)",
        "own": true,
        "primary": false,
        "parallel": true,
        "notes": "Claude Code harness; parallel test-time compute with an internal scoring model",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-18",
        "check": "human",
        "corroborated": false,
        "corroborated_on": null,
        "corroborated_by": [],
        "source_capture": "https://frontierlog.ai/captures/4e6f6af599737101.png"
      },
      {
        "id": "claude-opus-4:terminal-bench-1-0:default:3",
        "model": "claude-opus-4",
        "model_name": "Claude Opus 4",
        "model_date": "2025-05-22",
        "event": "https://frontierlog.ai/event/2025-05-22-claude-4",
        "benchmark": "terminal-bench-1-0",
        "benchmark_name": "Terminal-Bench 1.0",
        "setting": "default",
        "setting_name": "",
        "score": 39.2,
        "unit": "%",
        "printed": "39.2%",
        "reported_by": "anthropic",
        "run_by": null,
        "label": "Opus 4",
        "source": "https://www.anthropic.com/news/claude-4",
        "source_title": "Introducing Claude 4",
        "source_date": "2025-05-22",
        "where": "methodology footnote 2 under the main table",
        "own": true,
        "primary": false,
        "parallel": false,
        "notes": "the same agent as the other labs' models, not Claude Code; no extended thinking",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-18",
        "check": "human",
        "corroborated": false,
        "corroborated_on": null,
        "corroborated_by": []
      },
      {
        "id": "claude-opus-4:aime-2025:no-tools:1",
        "model": "claude-opus-4",
        "model_name": "Claude Opus 4",
        "model_date": "2025-05-22",
        "event": "https://frontierlog.ai/event/2025-05-22-claude-4",
        "benchmark": "aime-2025",
        "benchmark_name": "AIME 2025",
        "setting": "no-tools",
        "setting_name": "No tools",
        "score": 75.5,
        "unit": "%",
        "printed": "75.5%",
        "reported_by": "anthropic",
        "run_by": null,
        "label": "Claude Opus 4",
        "source": "https://www.anthropic.com/news/claude-4",
        "source_title": "Introducing Claude 4",
        "source_date": "2025-05-22",
        "where": "main benchmark table, row 'High school math competition / AIME 2025', column 'Claude Opus 4', first number of the cell",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "extended thinking up to 64K tokens",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-18",
        "check": "human",
        "corroborated": true,
        "corroborated_on": "2026-10-19",
        "corroborated_by": [
          {
            "url": "https://www.minimax.io/news/minimaxm1",
            "name": "MiniMax-M1",
            "title": "MiniMax-M1, the World's First Open-Source, Large-Scale, Hybrid-Attention Reasoning Model",
            "lab": "minimax",
            "date": "2025-06-16",
            "printed": "75.5",
            "where": "Full benchmark table image; row 'AIME 2025'; column 'Claude 4 Opus (64K thinking)'"
          }
        ],
        "source_capture": "https://frontierlog.ai/captures/4e6f6af599737101.png"
      },
      {
        "id": "claude-opus-4:aime-2025:no-tools:2",
        "model": "claude-opus-4",
        "model_name": "Claude Opus 4",
        "model_date": "2025-05-22",
        "event": "https://frontierlog.ai/event/2025-05-22-claude-4",
        "benchmark": "aime-2025",
        "benchmark_name": "AIME 2025",
        "setting": "no-tools",
        "setting_name": "No tools",
        "score": 90.0,
        "unit": "%",
        "printed": "90.0%",
        "reported_by": "anthropic",
        "run_by": null,
        "label": "Claude Opus 4",
        "source": "https://www.anthropic.com/news/claude-4",
        "source_title": "Introducing Claude 4",
        "source_date": "2025-05-22",
        "where": "main benchmark table, row 'AIME 2025', column 'Claude Opus 4', second number of the cell (after the slash)",
        "own": true,
        "primary": false,
        "parallel": true,
        "notes": "extended thinking up to 64K tokens; parallel test-time compute with internal scoring",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-18",
        "check": "human",
        "corroborated": false,
        "corroborated_on": null,
        "corroborated_by": [],
        "source_capture": "https://frontierlog.ai/captures/4e6f6af599737101.png"
      },
      {
        "id": "claude-opus-4:aime-2025:no-tools:3",
        "model": "claude-opus-4",
        "model_name": "Claude Opus 4",
        "model_date": "2025-05-22",
        "event": "https://frontierlog.ai/event/2025-05-22-claude-4",
        "benchmark": "aime-2025",
        "benchmark_name": "AIME 2025",
        "setting": "no-tools",
        "setting_name": "No tools",
        "score": 33.9,
        "unit": "%",
        "printed": "33.9%",
        "reported_by": "anthropic",
        "run_by": null,
        "label": "Opus 4",
        "source": "https://www.anthropic.com/news/claude-4",
        "source_title": "Introducing Claude 4",
        "source_date": "2025-05-22",
        "where": "appendix 'Performance benchmark reporting', extended-thinking list, AIME line",
        "own": true,
        "primary": false,
        "parallel": false,
        "notes": "no extended thinking",
        "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-max",
            "name": "Qwen3-Max",
            "title": "Qwen3-Max: Just Scale it",
            "lab": "alibaba",
            "date": "2025-09-24",
            "printed": "33.9",
            "where": "Qwen3-Max-Instruct bar chart (Qwen3-Max-Instruct.png), group AIME25, value label on the Claude Opus 4 Non-thinking bar"
          }
        ]
      }
    ]
  }
}