{
  "api_version": "1",
  "updated": "2026-10-26",
  "latest_event": "2026-10-24",
  "rights": "© Frontier Log, all rights reserved.",
  "model": {
    "slug": "claude-sonnet-5-5",
    "name": "Claude Sonnet 5.5",
    "lab": "anthropic",
    "date": "2026-09-28",
    "source": "https://www.anthropic.com/claude-sonnet-5-5",
    "source_title": "Introducing Claude Sonnet 5.5",
    "flagship": false,
    "event": "https://frontierlog.ai/event/2026-09-28-claude-sonnet-5-5",
    "releases": [],
    "variant_of": null,
    "made_by": "Anthropic",
    "url": "https://frontierlog.ai/model/claude-sonnet-5-5",
    "results": [
      {
        "id": "claude-sonnet-5-5:hle:no-tools:1",
        "model": "claude-sonnet-5-5",
        "model_name": "Claude Sonnet 5.5",
        "model_date": "2026-09-28",
        "event": "https://frontierlog.ai/event/2026-09-28-claude-sonnet-5-5",
        "benchmark": "hle",
        "benchmark_name": "Humanity's Last Exam",
        "setting": "no-tools",
        "setting_name": "No tools",
        "score": 56.9,
        "unit": "%",
        "printed": "56.9",
        "reported_by": "anthropic",
        "run_by": null,
        "label": "Claude Sonnet 5.5",
        "source": "https://www.anthropic.com/claude-sonnet-5-5-system-card",
        "source_title": "Claude Sonnet 5.5 System Card",
        "source_date": "2026-09-28",
        "where": "Section 8.1 evaluation summary table, row 'Humanity's Last Exam, No tools', column 'Claude Sonnet 5.5'",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "",
        "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-haiku-5-5-system-card",
            "name": "Claude Haiku 5.5 card",
            "title": "Claude Haiku 5.5 System Card",
            "lab": "anthropic",
            "date": "2026-10-07",
            "printed": "56.9",
            "where": "Table 8.1.A (p.111), row Humanity's Last Exam / No tools, column Sonnet 5.5"
          },
          {
            "url": "https://www.anthropic.com/claude-haiku-5-5",
            "name": "Claude Haiku 5.5",
            "title": "Introducing Claude Haiku 5.5",
            "lab": "anthropic",
            "date": "2026-10-07",
            "printed": "56.9%",
            "where": "HTML benchmark table 'Here's how Claude Haiku 5.5 performs', row Humanity's Last Exam (no tools), column Sonnet 5.5 (for reference)"
          }
        ],
        "source_capture": "https://frontierlog.ai/captures/53b68efd4b35b2e3.png"
      },
      {
        "id": "claude-sonnet-5-5:swe-bench-pro:default:1",
        "model": "claude-sonnet-5-5",
        "model_name": "Claude Sonnet 5.5",
        "model_date": "2026-09-28",
        "event": "https://frontierlog.ai/event/2026-09-28-claude-sonnet-5-5",
        "benchmark": "swe-bench-pro",
        "benchmark_name": "SWE-bench Pro",
        "setting": "default",
        "setting_name": "",
        "score": 81.3,
        "unit": "%",
        "printed": "81.3",
        "reported_by": "anthropic",
        "run_by": null,
        "label": "Claude Sonnet 5.5",
        "source": "https://www.anthropic.com/claude-sonnet-5-5-system-card",
        "source_title": "Claude Sonnet 5.5 System Card",
        "source_date": "2026-09-28",
        "where": "Section 8.1 Evaluation summary, Table 8.1.A, row 'SWE-Bench Pro', column 'Claude Sonnet 5.5'",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "adaptive thinking, max effort; 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": "2026-10-17",
        "check": "human",
        "corroborated": true,
        "corroborated_on": "2026-10-19",
        "corroborated_by": [
          {
            "url": "https://www.anthropic.com/claude-haiku-5-5-system-card",
            "name": "Claude Haiku 5.5 card",
            "title": "Claude Haiku 5.5 System Card",
            "lab": "anthropic",
            "date": "2026-10-07",
            "printed": "81.3",
            "where": "Table 8.1.A (p.111), row SWE-Bench Pro, column Sonnet 5.5"
          }
        ],
        "source_capture": "https://frontierlog.ai/captures/53b68efd4b35b2e3.png"
      },
      {
        "id": "claude-sonnet-5-5:deepswe-1-1:default:1",
        "model": "claude-sonnet-5-5",
        "model_name": "Claude Sonnet 5.5",
        "model_date": "2026-09-28",
        "event": "https://frontierlog.ai/event/2026-09-28-claude-sonnet-5-5",
        "benchmark": "deepswe-1-1",
        "benchmark_name": "DeepSWE v1.1",
        "setting": "default",
        "setting_name": "",
        "score": 71.0,
        "unit": "%",
        "printed": "71.0%",
        "reported_by": "anthropic",
        "run_by": null,
        "label": "Sonnet 5.5",
        "source": "https://www.anthropic.com/claude-sonnet-5-5-system-card",
        "source_title": "Claude Sonnet 5.5 System Card",
        "source_date": "2026-09-28",
        "where": "Section 8.3 'DeepSWE v1.1', paragraph",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "average of 5 trials",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-17",
        "check": "human",
        "corroborated": false,
        "corroborated_on": null,
        "corroborated_by": []
      },
      {
        "id": "claude-sonnet-5-5:terminal-bench-4-0:default:1",
        "model": "claude-sonnet-5-5",
        "model_name": "Claude Sonnet 5.5",
        "model_date": "2026-09-28",
        "event": "https://frontierlog.ai/event/2026-09-28-claude-sonnet-5-5",
        "benchmark": "terminal-bench-4-0",
        "benchmark_name": "Terminal-Bench 4.0",
        "setting": "default",
        "setting_name": "",
        "score": 70.6,
        "unit": "%",
        "printed": "70.6%",
        "reported_by": "anthropic",
        "run_by": null,
        "label": "Sonnet 5.5",
        "source": "https://www.anthropic.com/claude-sonnet-5-5",
        "source_title": "Introducing Claude Sonnet 5.5",
        "source_date": "2026-09-28",
        "where": "benchmark table, row 'Terminal-Bench 4.0', column 'Sonnet 5.5'",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "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://www.anthropic.com/claude-haiku-5-5-system-card",
            "name": "Claude Haiku 5.5 card",
            "title": "Claude Haiku 5.5 System Card",
            "lab": "anthropic",
            "date": "2026-10-07",
            "printed": "70.6%",
            "where": "Section 8.4 (p.115-116) sentence: \"Claude Sonnet 5.5 scored 70.6%\""
          },
          {
            "url": "https://www.anthropic.com/claude-sonnet-5-5-system-card",
            "name": "Claude Sonnet 5.5 card",
            "title": "Claude Sonnet 5.5 System Card",
            "lab": "anthropic",
            "date": "2026-09-28",
            "printed": "70.6%",
            "where": "Section 8.5 (p.113) sentence: \"Claude Sonnet 5.5 scored 70.6% on Terminal-Bench 4.0 with safeguards enabled\" (max effort)"
          },
          {
            "url": "https://www.anthropic.com/claude-haiku-5-5",
            "name": "Claude Haiku 5.5",
            "title": "Introducing Claude Haiku 5.5",
            "lab": "anthropic",
            "date": "2026-10-07",
            "printed": "70.6%",
            "where": "HTML benchmark table 'Here's how Claude Haiku 5.5 performs', row Terminal-Bench 4.0, column Sonnet 5.5 (for reference)"
          }
        ],
        "source_capture": "https://frontierlog.ai/captures/0a2dc9377a66cc01.png"
      },
      {
        "id": "claude-sonnet-5-5:hle:with-tools:1",
        "model": "claude-sonnet-5-5",
        "model_name": "Claude Sonnet 5.5",
        "model_date": "2026-09-28",
        "event": "https://frontierlog.ai/event/2026-09-28-claude-sonnet-5-5",
        "benchmark": "hle",
        "benchmark_name": "Humanity's Last Exam",
        "setting": "with-tools",
        "setting_name": "With tools",
        "score": 64.5,
        "unit": "%",
        "printed": "64.5%",
        "reported_by": "anthropic",
        "run_by": null,
        "label": "Sonnet 5.5",
        "source": "https://www.anthropic.com/claude-sonnet-5-5",
        "source_title": "Introducing Claude Sonnet 5.5",
        "source_date": "2026-09-28",
        "where": "benchmark table (HTML), row 'Multidisciplinary reasoning, Humanity's Last Exam', column 'Sonnet 5.5', 'with tools'",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-19",
        "check": "human",
        "corroborated": true,
        "corroborated_on": "2026-10-19",
        "corroborated_by": [
          {
            "url": "https://www.anthropic.com/claude-sonnet-5-5-system-card",
            "name": "Claude Sonnet 5.5 card",
            "title": "Claude Sonnet 5.5 System Card",
            "lab": "anthropic",
            "date": "2026-09-28",
            "printed": "64.5",
            "where": "Table 8.1.A, Humanity's Last Exam With tools row, Claude Sonnet 5.5 column"
          },
          {
            "url": "https://www.anthropic.com/claude-haiku-5-5",
            "name": "Claude Haiku 5.5",
            "title": "Introducing Claude Haiku 5.5",
            "lab": "anthropic",
            "date": "2026-10-07",
            "printed": "64.5%",
            "where": "benchmark table, Humanity's Last Exam with tools, Sonnet 5.5 (for reference) column"
          },
          {
            "url": "https://www.anthropic.com/claude-haiku-5-5-system-card",
            "name": "Claude Haiku 5.5 card",
            "title": "Claude Haiku 5.5 System Card",
            "lab": "anthropic",
            "date": "2026-10-07",
            "printed": "64.5",
            "where": "Table 8.1.A, Humanity's Last Exam With tools row, Sonnet 5.5 column"
          }
        ],
        "source_capture": "https://frontierlog.ai/captures/0a2dc9377a66cc01.png"
      },
      {
        "id": "claude-sonnet-5-5:terminal-bench-science-0-1:default:1",
        "model": "claude-sonnet-5-5",
        "model_name": "Claude Sonnet 5.5",
        "model_date": "2026-09-28",
        "event": "https://frontierlog.ai/event/2026-09-28-claude-sonnet-5-5",
        "benchmark": "terminal-bench-science-0-1",
        "benchmark_name": "Terminal-Bench Science 0.1",
        "setting": "default",
        "setting_name": "",
        "score": 59.9,
        "unit": "%",
        "printed": "59.9%",
        "reported_by": "anthropic",
        "run_by": null,
        "label": "Claude Sonnet 5.5",
        "source": "https://www.anthropic.com/claude-sonnet-5-5-system-card",
        "source_title": "Claude Sonnet 5.5 System Card",
        "source_date": "2026-09-28",
        "where": "Section 8.6 'Terminal-Bench-Science 0.1', first paragraph after the task description",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "Claude Code in --bare mode; max effort; standard error ±4.7 points",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-19",
        "check": "human",
        "corroborated": true,
        "corroborated_on": "2026-10-19",
        "corroborated_by": [
          {
            "url": "https://www.anthropic.com/claude-haiku-5-5-system-card",
            "name": "Claude Haiku 5.5 card",
            "title": "Claude Haiku 5.5 System Card",
            "lab": "anthropic",
            "date": "2026-10-07",
            "printed": "59.9%",
            "where": "section 8.5 sentence: 'Claude Sonnet 5.5 scored 59.9%' on Terminal-Bench-Science"
          }
        ]
      },
      {
        "id": "claude-sonnet-5-5:automationbench:default:1",
        "model": "claude-sonnet-5-5",
        "model_name": "Claude Sonnet 5.5",
        "model_date": "2026-09-28",
        "event": "https://frontierlog.ai/event/2026-09-28-claude-sonnet-5-5",
        "benchmark": "automationbench",
        "benchmark_name": "AutomationBench",
        "setting": "default",
        "setting_name": "",
        "score": 44.7,
        "unit": "%",
        "printed": "44.7",
        "reported_by": "anthropic",
        "run_by": null,
        "label": "Claude Sonnet 5.5",
        "source": "https://www.anthropic.com/claude-sonnet-5-5-system-card",
        "source_title": "Claude Sonnet 5.5 System Card",
        "source_date": "2026-09-28",
        "where": "Section 8.1, Table 8.1.A, row 'AutomationBench', column 'Claude Sonnet 5.5'",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "run by Zapier; private set, release 1.0.6; max effort; default fallbacks on",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-20",
        "check": "human",
        "corroborated": false,
        "corroborated_on": null,
        "corroborated_by": [],
        "source_capture": "https://frontierlog.ai/captures/53b68efd4b35b2e3.png"
      },
      {
        "id": "claude-sonnet-5-5:frontiercode-1-1:main:1",
        "model": "claude-sonnet-5-5",
        "model_name": "Claude Sonnet 5.5",
        "model_date": "2026-09-28",
        "event": "https://frontierlog.ai/event/2026-09-28-claude-sonnet-5-5",
        "benchmark": "frontiercode-1-1",
        "benchmark_name": "FrontierCode 1.1",
        "setting": "main",
        "setting_name": "Main set",
        "score": 46.2,
        "unit": "%",
        "printed": "46.2%",
        "reported_by": "anthropic",
        "run_by": null,
        "label": "Sonnet 5.5 (Max)",
        "source": "https://www.anthropic.com/claude-sonnet-5-5",
        "source_title": "Introducing Claude Sonnet 5.5",
        "source_date": "2026-09-28",
        "where": "benchmark table (HTML), row 'Agentic coding, FrontierCode 1.1 (Main)', column 'Sonnet 5.5', the 'Max' value",
        "own": true,
        "primary": false,
        "parallel": false,
        "notes": "max effort; run by Cognition in Claude Code",
        "caveat": "",
        "human_verified": false,
        "human_verified_on": null,
        "check": "corroborated",
        "corroborated": true,
        "corroborated_on": "2026-10-19",
        "corroborated_by": [
          {
            "url": "https://www.anthropic.com/claude-sonnet-5-5-system-card",
            "name": "Claude Sonnet 5.5 card",
            "title": "Claude Sonnet 5.5 System Card",
            "lab": "anthropic",
            "date": "2026-09-28",
            "printed": "46.2",
            "where": "Table 8.1.A, row FrontierCode v1.1 (Main), column Claude Sonnet 5.5 (max effort; 8.4 text 'At max effort, Sonnet 5.5 scores 46.2% on Main')"
          },
          {
            "url": "https://www.anthropic.com/claude-haiku-5-5-system-card",
            "name": "Claude Haiku 5.5 card",
            "title": "Claude Haiku 5.5 System Card",
            "lab": "anthropic",
            "date": "2026-10-07",
            "printed": "46.2",
            "where": "Table 8.1.A, row FrontierCode 1.1 (Main), column Sonnet 5.5, '(max)'"
          }
        ],
        "source_capture": "https://frontierlog.ai/captures/0a2dc9377a66cc01.png"
      },
      {
        "id": "claude-sonnet-5-5:frontiercode-1-1:main:2",
        "model": "claude-sonnet-5-5",
        "model_name": "Claude Sonnet 5.5",
        "model_date": "2026-09-28",
        "event": "https://frontierlog.ai/event/2026-09-28-claude-sonnet-5-5",
        "benchmark": "frontiercode-1-1",
        "benchmark_name": "FrontierCode 1.1",
        "setting": "main",
        "setting_name": "Main set",
        "score": 52.1,
        "unit": "%",
        "printed": "52.1%",
        "reported_by": "anthropic",
        "run_by": null,
        "label": "Sonnet 5.5 (Xhigh)",
        "source": "https://www.anthropic.com/claude-sonnet-5-5",
        "source_title": "Introducing Claude Sonnet 5.5",
        "source_date": "2026-09-28",
        "where": "benchmark table (HTML), row 'Agentic coding, FrontierCode 1.1 (Main)', column 'Sonnet 5.5', the 'Xhigh' value",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "xhigh effort; run by Cognition in Claude Code",
        "caveat": "",
        "human_verified": false,
        "human_verified_on": null,
        "check": "corroborated",
        "corroborated": true,
        "corroborated_on": "2026-10-19",
        "corroborated_by": [
          {
            "url": "https://www.anthropic.com/claude-sonnet-5-5-system-card",
            "name": "Claude Sonnet 5.5 card",
            "title": "Claude Sonnet 5.5 System Card",
            "lab": "anthropic",
            "date": "2026-09-28",
            "printed": "52.1%",
            "where": "Section 8.4 text: 'At xhigh effort, Claude Sonnet 5.5 scores 52.1% on Main'"
          },
          {
            "url": "https://www.anthropic.com/claude-haiku-5-5-system-card",
            "name": "Claude Haiku 5.5 card",
            "title": "Claude Haiku 5.5 System Card",
            "lab": "anthropic",
            "date": "2026-10-07",
            "printed": "52.1",
            "where": "Table 8.1.A, row FrontierCode 1.1 (Main), column Sonnet 5.5, '(xhigh)'"
          },
          {
            "url": "https://www.anthropic.com/claude-haiku-5-5",
            "name": "Claude Haiku 5.5",
            "title": "Introducing Claude Haiku 5.5",
            "lab": "anthropic",
            "date": "2026-10-07",
            "printed": "52.1%",
            "where": "Benchmark table, row Agentic coding FrontierCode 1.1 (Main), column Sonnet 5.5 (For reference), 'Xhigh'"
          }
        ],
        "source_capture": "https://frontierlog.ai/captures/0a2dc9377a66cc01.png"
      },
      {
        "id": "claude-sonnet-5-5:frontiercode-1-1:extended:1",
        "model": "claude-sonnet-5-5",
        "model_name": "Claude Sonnet 5.5",
        "model_date": "2026-09-28",
        "event": "https://frontierlog.ai/event/2026-09-28-claude-sonnet-5-5",
        "benchmark": "frontiercode-1-1",
        "benchmark_name": "FrontierCode 1.1",
        "setting": "extended",
        "setting_name": "Extended set",
        "score": 59.1,
        "unit": "%",
        "printed": "59.1%",
        "reported_by": "anthropic",
        "run_by": null,
        "label": "Claude Sonnet 5.5",
        "source": "https://www.anthropic.com/claude-sonnet-5-5-system-card",
        "source_title": "Claude Sonnet 5.5 System Card",
        "source_date": "2026-09-28",
        "where": "Section 8.4 'FrontierCode', paragraph 'At xhigh effort ...', the max-effort Extended figure",
        "own": true,
        "primary": false,
        "parallel": false,
        "notes": "max effort; run by Cognition in Claude Code",
        "caveat": "",
        "human_verified": false,
        "human_verified_on": null,
        "check": "corroborated",
        "corroborated": true,
        "corroborated_on": "2026-10-19",
        "corroborated_by": [
          {
            "url": "https://www.anthropic.com/claude-haiku-5-5-system-card",
            "name": "Claude Haiku 5.5 card",
            "title": "Claude Haiku 5.5 System Card",
            "lab": "anthropic",
            "date": "2026-10-07",
            "printed": "59.1%",
            "where": "Section 8.3 text: 'Claude Sonnet 5.5 scored 46.2% on Main and 59.1% on Extended at max effort'"
          }
        ]
      },
      {
        "id": "claude-sonnet-5-5:frontiercode-1-1:extended:2",
        "model": "claude-sonnet-5-5",
        "model_name": "Claude Sonnet 5.5",
        "model_date": "2026-09-28",
        "event": "https://frontierlog.ai/event/2026-09-28-claude-sonnet-5-5",
        "benchmark": "frontiercode-1-1",
        "benchmark_name": "FrontierCode 1.1",
        "setting": "extended",
        "setting_name": "Extended set",
        "score": 64.4,
        "unit": "%",
        "printed": "64.4%",
        "reported_by": "anthropic",
        "run_by": null,
        "label": "Claude Sonnet 5.5",
        "source": "https://www.anthropic.com/claude-sonnet-5-5-system-card",
        "source_title": "Claude Sonnet 5.5 System Card",
        "source_date": "2026-09-28",
        "where": "Section 8.4 'FrontierCode', paragraph 'At xhigh effort ...', the xhigh-effort Extended figure",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "xhigh effort; run by Cognition in Claude Code",
        "caveat": "",
        "human_verified": false,
        "human_verified_on": null,
        "check": "corroborated",
        "corroborated": true,
        "corroborated_on": "2026-10-19",
        "corroborated_by": [
          {
            "url": "https://www.anthropic.com/claude-haiku-5-5-system-card",
            "name": "Claude Haiku 5.5 card",
            "title": "Claude Haiku 5.5 System Card",
            "lab": "anthropic",
            "date": "2026-10-07",
            "printed": "64.4%",
            "where": "Section 8.3 text: 'Its best scores were 52.1% and 64.4%, at xhigh' (Sonnet 5.5, Main and Extended)"
          }
        ]
      },
      {
        "id": "claude-sonnet-5-5:osworld-2-1:partial:1",
        "model": "claude-sonnet-5-5",
        "model_name": "Claude Sonnet 5.5",
        "model_date": "2026-09-28",
        "event": "https://frontierlog.ai/event/2026-09-28-claude-sonnet-5-5",
        "benchmark": "osworld-2-1",
        "benchmark_name": "OSWorld 2.1",
        "setting": "partial",
        "setting_name": "Full set, partial score",
        "score": 80.1,
        "unit": "%",
        "printed": "80.1%",
        "reported_by": "anthropic",
        "run_by": null,
        "label": "Sonnet 5.5",
        "source": "https://www.anthropic.com/claude-sonnet-5-5",
        "source_title": "Introducing Claude Sonnet 5.5",
        "source_date": "2026-09-28",
        "where": "benchmark table (HTML), row 'Computer use / OSWorld 2.1', column 'Sonnet 5.5', the value tagged 'partial'",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "max effort; mean of 5 runs; September 10, 2026 task files",
        "caveat": "",
        "human_verified": false,
        "human_verified_on": null,
        "check": "corroborated",
        "corroborated": true,
        "corroborated_on": "2026-10-20",
        "corroborated_by": [
          {
            "url": "https://www.anthropic.com/claude-sonnet-5-5-system-card",
            "name": "Claude Sonnet 5.5 card",
            "title": "Claude Sonnet 5.5 System Card",
            "lab": "anthropic",
            "date": "2026-09-28",
            "printed": "80.1%",
            "where": "section 8.13.3 OSWorld 2.1 (p. 130), sentence: 'Claude Sonnet 5.5 achieved a partial score of 80.1% and a strict pass rate of 43.5%'; also 8.1 Evaluation summary, row 'OSWorld 2.1', column Claude Sonnet 5.5: '80.1'"
          }
        ],
        "source_capture": "https://frontierlog.ai/captures/0a2dc9377a66cc01.png"
      },
      {
        "id": "claude-sonnet-5-5:osworld-2-1:strict:1",
        "model": "claude-sonnet-5-5",
        "model_name": "Claude Sonnet 5.5",
        "model_date": "2026-09-28",
        "event": "https://frontierlog.ai/event/2026-09-28-claude-sonnet-5-5",
        "benchmark": "osworld-2-1",
        "benchmark_name": "OSWorld 2.1",
        "setting": "strict",
        "setting_name": "Full set, strict pass rate",
        "score": 43.5,
        "unit": "%",
        "printed": "43.5%",
        "reported_by": "anthropic",
        "run_by": null,
        "label": "Claude Sonnet 5.5",
        "source": "https://www.anthropic.com/claude-sonnet-5-5-system-card",
        "source_title": "Claude Sonnet 5.5 System Card",
        "source_date": "2026-09-28",
        "where": "section 8.13.3 OSWorld 2.1, text (PDF page 130), the Claude Sonnet 5.5 strict pass rate",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "max effort; mean of 5 runs; September 10, 2026 task files",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-20",
        "check": "human",
        "corroborated": false,
        "corroborated_on": null,
        "corroborated_by": [],
        "source_capture": "https://frontierlog.ai/captures/ed9b8a21a7f60448.png"
      },
      {
        "id": "claude-sonnet-5-5:osworld-2-1:offline-partial:1",
        "model": "claude-sonnet-5-5",
        "model_name": "Claude Sonnet 5.5",
        "model_date": "2026-09-28",
        "event": "https://frontierlog.ai/event/2026-09-28-claude-sonnet-5-5",
        "benchmark": "osworld-2-1",
        "benchmark_name": "OSWorld 2.1",
        "setting": "offline-partial",
        "setting_name": "Offline set, partial score",
        "score": 83.9,
        "unit": "%",
        "printed": "83.9%",
        "reported_by": "anthropic",
        "run_by": null,
        "label": "Sonnet 5.5",
        "source": "https://www.anthropic.com/claude-haiku-5-5",
        "source_title": "Introducing Claude Haiku 5.5",
        "source_date": "2026-10-07",
        "where": "benchmark table (HTML), row 'Computer use / OSWorld 2.1', column 'Sonnet 5.5', the value tagged 'Offline subset'",
        "own": false,
        "primary": true,
        "parallel": false,
        "notes": "82-task offline subset, no internet access; max effort; mean of 5 attempts",
        "caveat": "",
        "human_verified": false,
        "human_verified_on": null,
        "check": "corroborated",
        "corroborated": true,
        "corroborated_on": "2026-10-20",
        "corroborated_by": [
          {
            "url": "https://www.anthropic.com/claude-haiku-5-5-system-card",
            "name": "Claude Haiku 5.5 card",
            "title": "Claude Haiku 5.5 System Card",
            "lab": "anthropic",
            "date": "2026-10-07",
            "printed": "83.9%",
            "where": "section 8.9.3 OSWorld 2.1 offline subset (p. 128), sentence: 'Claude Sonnet 5.5 scored 83.9% partial and 48.8% strict pass'; also 8.1 Evaluation summary, row 'OSWorld 2.1 (offline subset)', column Sonnet 5.5: '83.9'"
          }
        ],
        "source_capture": "https://frontierlog.ai/captures/119b9e1f2a7a6ddd.png"
      },
      {
        "id": "claude-sonnet-5-5:osworld-2-1:offline-strict:1",
        "model": "claude-sonnet-5-5",
        "model_name": "Claude Sonnet 5.5",
        "model_date": "2026-09-28",
        "event": "https://frontierlog.ai/event/2026-09-28-claude-sonnet-5-5",
        "benchmark": "osworld-2-1",
        "benchmark_name": "OSWorld 2.1",
        "setting": "offline-strict",
        "setting_name": "Offline set, strict pass rate",
        "score": 48.8,
        "unit": "%",
        "printed": "48.8%",
        "reported_by": "anthropic",
        "run_by": null,
        "label": "Claude Sonnet 5.5",
        "source": "https://www.anthropic.com/claude-haiku-5-5-system-card",
        "source_title": "Claude Haiku 5.5 System Card",
        "source_date": "2026-10-07",
        "where": "section 8.9.3 OSWorld 2.1 offline subset, text (PDF page 128), the Claude Sonnet 5.5 strict pass rate",
        "own": false,
        "primary": true,
        "parallel": false,
        "notes": "82-task offline subset, no internet access; max effort; mean of 5 attempts",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-20",
        "check": "human",
        "corroborated": false,
        "corroborated_on": null,
        "corroborated_by": [],
        "source_capture": "https://frontierlog.ai/captures/78b2fca2e8e996c4.png"
      }
    ]
  }
}