{
  "api_version": "1",
  "updated": "2026-10-17",
  "latest_event": "2026-10-17",
  "rights": "© Frontier Log, all rights reserved.",
  "benchmark": {
    "slug": "swe-bench-pro",
    "name": "SWE-bench Pro",
    "family": "swe-bench",
    "unit": "%",
    "higher_is_better": true,
    "status": "active",
    "benchmark_url": "https://scale.com/leaderboard/swe_bench_pro_public",
    "introduced": "2025-09-19",
    "creator": "Scale AI",
    "introduced_source": "https://scale.com/blog/swe-bench-pro",
    "settings": [
      {
        "setting": "default",
        "name": "",
        "url": "https://frontierlog.ai/benchmark/swe-bench-pro",
        "results": 28
      }
    ]
  },
  "total": 28,
  "results": [
    {
      "id": "gemini-3-pro:swe-bench-pro:default:1",
      "model": "gemini-3-pro",
      "model_name": "Gemini 3 Pro",
      "model_date": "2025-11-18",
      "event": "https://frontierlog.ai/event/2025-11-18-gemini-3-pro",
      "benchmark": "swe-bench-pro",
      "benchmark_name": "SWE-bench Pro",
      "setting": "default",
      "setting_name": "",
      "score": 43.3,
      "unit": "%",
      "printed": "43.3%",
      "reported_by": "google",
      "label": "Gemini 3 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 Pro Thinking (High)'",
      "own": false,
      "primary": true,
      "parallel": false,
      "notes": "single attempt, Thinking (High)",
      "caveat": "",
      "human_verified": true,
      "human_verified_on": "2026-10-17",
      "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": "gpt-5-2:swe-bench-pro:default:1",
      "model": "gpt-5-2",
      "model_name": "GPT-5.2",
      "model_date": "2025-12-11",
      "event": "https://frontierlog.ai/event/2025-12-11-gpt-5-2",
      "benchmark": "swe-bench-pro",
      "benchmark_name": "SWE-bench Pro",
      "setting": "default",
      "setting_name": "",
      "score": 55.6,
      "unit": "%",
      "printed": "55.6%",
      "reported_by": "openai",
      "label": "GPT-5.2 Thinking",
      "source": "https://openai.com/index/introducing-gpt-5-2",
      "source_title": "Introducing GPT-5.2",
      "source_date": "2025-12-11",
      "where": "coding table (columns GPT-5.2 Thinking, GPT-5.2 Pro, GPT-5.1 Thinking), row 'SWE-Bench Pro, Public', column 'GPT-5.2 Thinking'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "",
      "caveat": "",
      "human_verified": true,
      "human_verified_on": "2026-10-17"
    },
    {
      "id": "kimi-k2-5:swe-bench-pro:default:1",
      "model": "kimi-k2-5",
      "model_name": "Kimi K2.5",
      "model_date": "2026-01-27",
      "event": "https://frontierlog.ai/event/2026-01-27-kimi-k2-5",
      "benchmark": "swe-bench-pro",
      "benchmark_name": "SWE-bench Pro",
      "setting": "default",
      "setting_name": "",
      "score": 50.7,
      "unit": "%",
      "printed": "50.7",
      "reported_by": "moonshot",
      "label": "Kimi K2.5 (Thinking)",
      "source": "https://www.kimi.ai/blog/kimi-k2-5",
      "source_title": "Kimi K2.5: Visual Agentic Intelligence",
      "source_date": "2026-01-27",
      "where": "Appendix 'Benchmark table', row 'SWE-Bench Pro' (Coding), column 'Kimi K2.5 (Thinking)' (read from the page's embedded benchmark JSON and the rendered table)",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "in-house minimal-tool framework, non-thinking mode, average of 5 runs",
      "caveat": "The source doesn't say whether this is the public or the commercial set",
      "human_verified": true,
      "human_verified_on": "2026-10-17"
    },
    {
      "id": "gpt-5-3-codex:swe-bench-pro:default:1",
      "model": "gpt-5-3-codex",
      "model_name": "GPT-5.3-Codex",
      "model_date": "2026-02-05",
      "event": "https://frontierlog.ai/event/2026-02-05-gpt-5-3-codex",
      "benchmark": "swe-bench-pro",
      "benchmark_name": "SWE-bench Pro",
      "setting": "default",
      "setting_name": "",
      "score": 56.8,
      "unit": "%",
      "printed": "56.8%",
      "reported_by": "openai",
      "label": "GPT‑5.3‑Codex (xhigh)",
      "source": "https://openai.com/index/introducing-gpt-5-3-codex",
      "source_title": "Introducing GPT-5.3-Codex",
      "source_date": "2026-02-05",
      "where": "eval table, row 'SWE-Bench Pro (Public)', column 'GPT‑5.3‑Codex (xhigh)'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "xhigh reasoning effort",
      "caveat": "",
      "human_verified": true,
      "human_verified_on": "2026-10-17"
    },
    {
      "id": "minimax-m2-5:swe-bench-pro:default:1",
      "model": "minimax-m2-5",
      "model_name": "MiniMax M2.5",
      "model_date": "2026-02-12",
      "event": "https://frontierlog.ai/event/2026-02-12-minimax-m2-5",
      "benchmark": "swe-bench-pro",
      "benchmark_name": "SWE-bench Pro",
      "setting": "default",
      "setting_name": "",
      "score": 55.4,
      "unit": "%",
      "printed": "55.4",
      "reported_by": "minimax",
      "label": "MiniMax M2.5",
      "source": "https://www.minimax.io/news/minimax-m25",
      "source_title": "MiniMax M2.5: Built for Real-World Productivity.",
      "source_date": "2026-02-12",
      "where": "first benchmark chart image (top of the post, eight panels), panel 'SWE-Bench Pro', bar 'MiniMax M2.5' (leftmost), value label; the same label appears in the later coding chart image 54ddb070-….png",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "Claude Code scaffold, average of 4 runs",
      "caveat": "The source doesn't say whether this is the public or the commercial set",
      "human_verified": true,
      "human_verified_on": "2026-10-17",
      "source_image": "https://frontierlog.ai/sources/7eaf807b2996051e.png",
      "source_image_original": "https://file.cdn.minimax.io/public/97f76950-2c60-4a9b-bb96-228454afabe9.png"
    },
    {
      "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",
      "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",
      "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": "gpt-5-4:swe-bench-pro:default:1",
      "model": "gpt-5-4",
      "model_name": "GPT-5.4",
      "model_date": "2026-03-05",
      "event": "https://frontierlog.ai/event/2026-03-05-gpt-5-4",
      "benchmark": "swe-bench-pro",
      "benchmark_name": "SWE-bench Pro",
      "setting": "default",
      "setting_name": "",
      "score": 57.7,
      "unit": "%",
      "printed": "57.7%",
      "reported_by": "openai",
      "label": "GPT‑5.4",
      "source": "https://openai.com/index/introducing-gpt-5-4",
      "source_title": "Introducing GPT-5.4",
      "source_date": "2026-03-05",
      "where": "coding table (columns GPT‑5.4, GPT‑5.4 Pro, GPT‑5.3-Codex, …), row 'SWE-Bench Pro (Public)', column 'GPT‑5.4'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "",
      "caveat": "",
      "human_verified": true,
      "human_verified_on": "2026-10-17"
    },
    {
      "id": "claude-mythos-preview:swe-bench-pro:default:1",
      "model": "claude-mythos-preview",
      "model_name": "Claude Mythos Preview",
      "model_date": "2026-04-07",
      "event": "https://frontierlog.ai/event/2026-04-07-claude-mythos-preview",
      "benchmark": "swe-bench-pro",
      "benchmark_name": "SWE-bench Pro",
      "setting": "default",
      "setting_name": "",
      "score": 77.8,
      "unit": "%",
      "printed": "77.8%",
      "reported_by": "anthropic",
      "label": "Mythos Preview",
      "source": "https://www.anthropic.com/news/claude-opus-4-7",
      "source_title": "Introducing Claude Opus 4.7",
      "source_date": "2026-04-16",
      "where": "benchmark table (image), row 'Agentic coding, SWE-bench Pro', column 'Mythos Preview'",
      "own": false,
      "primary": true,
      "parallel": false,
      "notes": "",
      "caveat": "The source doesn't say whether this is the public or the commercial set",
      "human_verified": true,
      "human_verified_on": "2026-10-17",
      "source_image": "https://frontierlog.ai/sources/527cd462f89b6de4.png",
      "source_image_original": "https://www-cdn.anthropic.com/images/4zrzovbb/website/d434d15757c6abac1122af483617741776d5a114-2600x2638.png"
    },
    {
      "id": "claude-opus-4-7:swe-bench-pro:default:1",
      "model": "claude-opus-4-7",
      "model_name": "Claude Opus 4.7",
      "model_date": "2026-04-16",
      "event": "https://frontierlog.ai/event/2026-04-16-claude-opus-4-7",
      "benchmark": "swe-bench-pro",
      "benchmark_name": "SWE-bench Pro",
      "setting": "default",
      "setting_name": "",
      "score": 64.3,
      "unit": "%",
      "printed": "64.3%",
      "reported_by": "anthropic",
      "label": "Opus 4.7",
      "source": "https://www.anthropic.com/news/claude-opus-4-7",
      "source_title": "Introducing Claude Opus 4.7",
      "source_date": "2026-04-16",
      "where": "benchmark table (image), row 'Agentic coding, SWE-bench Pro', column 'Opus 4.7'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "memorization screens flag a subset of problems; the margin over Opus 4.6 holds without them",
      "caveat": "The source doesn't say whether this is the public or the commercial set",
      "human_verified": true,
      "human_verified_on": "2026-10-17",
      "source_image": "https://frontierlog.ai/sources/527cd462f89b6de4.png",
      "source_image_original": "https://www-cdn.anthropic.com/images/4zrzovbb/website/d434d15757c6abac1122af483617741776d5a114-2600x2638.png"
    },
    {
      "id": "kimi-k2-6:swe-bench-pro:default:1",
      "model": "kimi-k2-6",
      "model_name": "Kimi K2.6",
      "model_date": "2026-04-20",
      "event": "https://frontierlog.ai/event/2026-04-20-kimi-k2-6",
      "benchmark": "swe-bench-pro",
      "benchmark_name": "SWE-bench Pro",
      "setting": "default",
      "setting_name": "",
      "score": 58.6,
      "unit": "%",
      "printed": "58.6",
      "reported_by": "moonshot",
      "label": "Kimi K2.6",
      "source": "https://www.kimi.ai/blog/kimi-k2-6",
      "source_title": "Kimi K2.6: Advancing Open-Source Coding",
      "source_date": "2026-04-20",
      "where": "'Benchmark Table' near the end, row 'SWE-Bench Pro' (Coding), column 'Kimi K2.6' (embedded benchmark JSON; the same value is in the coding chart 'SWE-Bench Pro' near the top). Ignore the 'Model A/Model B' placeholder widget also embedded in the page",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "in-house framework adapted from SWE-agent, thinking mode, average of 10 runs",
      "caveat": "The source doesn't say whether this is the public or the commercial set",
      "human_verified": true,
      "human_verified_on": "2026-10-17"
    },
    {
      "id": "gpt-5-5:swe-bench-pro:default:1",
      "model": "gpt-5-5",
      "model_name": "GPT-5.5",
      "model_date": "2026-04-23",
      "event": "https://frontierlog.ai/event/2026-04-23-gpt-5-5",
      "benchmark": "swe-bench-pro",
      "benchmark_name": "SWE-bench Pro",
      "setting": "default",
      "setting_name": "",
      "score": 58.6,
      "unit": "%",
      "printed": "58.6%",
      "reported_by": "openai",
      "label": "GPT‑5.5",
      "source": "https://openai.com/index/introducing-gpt-5-5",
      "source_title": "Introducing GPT-5.5",
      "source_date": "2026-04-23",
      "where": "coding eval table, row 'SWE-Bench Pro (Public) *', column 'GPT‑5.5'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "",
      "caveat": "",
      "human_verified": true,
      "human_verified_on": "2026-10-17"
    },
    {
      "id": "deepseek-v4-pro:swe-bench-pro:default:1",
      "model": "deepseek-v4-pro",
      "model_name": "DeepSeek-V4-Pro",
      "model_date": "2026-04-24",
      "event": "https://frontierlog.ai/event/2026-04-24-deepseek-v4-preview",
      "benchmark": "swe-bench-pro",
      "benchmark_name": "SWE-bench Pro",
      "setting": "default",
      "setting_name": "",
      "score": 55.4,
      "unit": "%",
      "printed": "55.4",
      "reported_by": "deepseek",
      "label": "DS-V4-Pro Max",
      "source": "https://api-docs.deepseek.com/news/news260424/",
      "source_title": "DeepSeek V4 Preview Release",
      "source_date": "2026-04-24",
      "where": "benchmark table image, row 'SWE Pro (Resolved)' (Agentic), column 'DS-V4-Pro Max'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "Max reasoning effort",
      "caveat": "The source doesn't say whether this is the public or the commercial set",
      "human_verified": true,
      "human_verified_on": "2026-10-17",
      "source_image": "https://frontierlog.ai/sources/9226203de8e24b33.png",
      "source_image_original": "https://api-docs.deepseek.com/img/v4-benchmark-2.png"
    },
    {
      "id": "gemini-3-5-flash:swe-bench-pro:default:1",
      "model": "gemini-3-5-flash",
      "model_name": "Gemini 3.5 Flash",
      "model_date": "2026-05-19",
      "event": "https://frontierlog.ai/event/2026-05-19-gemini-3-5-flash",
      "benchmark": "swe-bench-pro",
      "benchmark_name": "SWE-bench Pro",
      "setting": "default",
      "setting_name": "",
      "score": 55.1,
      "unit": "%",
      "printed": "55.1%",
      "reported_by": "google",
      "label": "Gemini 3.5 Flash",
      "source": "https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-5/",
      "source_title": "Gemini 3.5: frontier intelligence with action",
      "source_date": "2026-05-19",
      "where": "benchmark table (image), Coding group, row 'SWE-Bench Pro (Public)', column 'Gemini 3.5 Flash'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "single attempt",
      "caveat": "",
      "human_verified": true,
      "human_verified_on": "2026-10-17",
      "source_image": "https://frontierlog.ai/sources/9db2ce79d7fae814.gif",
      "source_image_original": "https://storage.googleapis.com/gweb-uniblog-publish-prod/original_images/gemini-3-5__benchmarks__light.gif"
    },
    {
      "id": "claude-opus-4-8:swe-bench-pro:default:1",
      "model": "claude-opus-4-8",
      "model_name": "Claude Opus 4.8",
      "model_date": "2026-05-28",
      "event": "https://frontierlog.ai/event/2026-05-28-claude-opus-4-8",
      "benchmark": "swe-bench-pro",
      "benchmark_name": "SWE-bench Pro",
      "setting": "default",
      "setting_name": "",
      "score": 69.2,
      "unit": "%",
      "printed": "69.2%",
      "reported_by": "anthropic",
      "label": "Opus 4.8",
      "source": "https://www.anthropic.com/news/claude-opus-4-8",
      "source_title": "Introducing Claude Opus 4.8",
      "source_date": "2026-05-28",
      "where": "benchmark table (image), row 'Agentic coding, SWE-Bench Pro', column 'Opus 4.8'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "",
      "caveat": "The source doesn't say whether this is the public or the commercial set",
      "human_verified": true,
      "human_verified_on": "2026-10-17",
      "source_image": "https://frontierlog.ai/sources/7ad9b784f2579c8a.png",
      "source_image_original": "https://www-cdn.anthropic.com/images/4zrzovbb/website/a9007019094f217e98cb8261a2765d7646c01708-2600x1392.png"
    },
    {
      "id": "minimax-m3:swe-bench-pro:default:1",
      "model": "minimax-m3",
      "model_name": "MiniMax M3",
      "model_date": "2026-06-01",
      "event": "https://frontierlog.ai/event/2026-06-01-minimax-m3",
      "benchmark": "swe-bench-pro",
      "benchmark_name": "SWE-bench Pro",
      "setting": "default",
      "setting_name": "",
      "score": 59.0,
      "unit": "%",
      "printed": "59.0",
      "reported_by": "minimax",
      "label": "MiniMax M3",
      "source": "https://www.minimax.io/blog/minimax-m3",
      "source_title": "MiniMax M3: Frontier Coding, 1M Context, Native Multimodality — All in One Model",
      "source_date": "2026-06-01",
      "where": "benchmark table image, row 'SWE-Bench Pro' (Coding), column 'MiniMax M3' (the post also lists it in prose under the coding paragraph and in the SWE Bench Pro chart image)",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "Claude Code scaffold",
      "caveat": "The source doesn't say whether this is the public or the commercial set",
      "human_verified": true,
      "human_verified_on": "2026-10-17",
      "source_image": "https://frontierlog.ai/sources/5720f3ad06ea2ce9.jpg",
      "source_image_original": "https://filecdn.minimax.chat/public/img_v3_02128_b7726cd8-879a-4b7a-a9da-db4395ea597g-1780272508686.jpg"
    },
    {
      "id": "claude-fable-5:swe-bench-pro:default:1",
      "model": "claude-fable-5",
      "model_name": "Claude Fable 5",
      "model_date": "2026-06-09",
      "event": "https://frontierlog.ai/event/2026-06-09-claude-fable-5-mythos-5",
      "benchmark": "swe-bench-pro",
      "benchmark_name": "SWE-bench Pro",
      "setting": "default",
      "setting_name": "",
      "score": 80.0,
      "unit": "%",
      "printed": "80",
      "reported_by": "anthropic",
      "label": "Fable 5",
      "source": "https://www.anthropic.com/claude-fable-5-mythos-5-system-card",
      "source_title": "Claude Fable 5 and Claude Mythos 5 System Card",
      "source_date": "2026-06-09",
      "where": "Section 8.1 Evaluation summary, Table 8.1.A, row 'SWE-bench Pro', column 'Fable 5'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "adaptive thinking, max effort, average of 5 trials; with production safeguards, including fallback to Opus 4.8",
      "caveat": "The source doesn't say whether this is the public or the commercial set",
      "human_verified": true,
      "human_verified_on": "2026-10-17"
    },
    {
      "id": "claude-mythos-5:swe-bench-pro:default:1",
      "model": "claude-mythos-5",
      "model_name": "Claude Mythos 5",
      "model_date": "2026-06-09",
      "event": "https://frontierlog.ai/event/2026-06-09-claude-fable-5-mythos-5",
      "benchmark": "swe-bench-pro",
      "benchmark_name": "SWE-bench Pro",
      "setting": "default",
      "setting_name": "",
      "score": 80.3,
      "unit": "%",
      "printed": "80.3",
      "reported_by": "anthropic",
      "label": "Mythos 5",
      "source": "https://www.anthropic.com/claude-fable-5-mythos-5-system-card",
      "source_title": "Claude Fable 5 and Claude Mythos 5 System Card",
      "source_date": "2026-06-09",
      "where": "Section 8.1 Evaluation summary, Table 8.1.A, row 'SWE-bench Pro', column 'Mythos 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"
    },
    {
      "id": "claude-sonnet-5:swe-bench-pro:default:1",
      "model": "claude-sonnet-5",
      "model_name": "Claude Sonnet 5",
      "model_date": "2026-06-30",
      "event": "https://frontierlog.ai/event/2026-06-30-claude-sonnet-5",
      "benchmark": "swe-bench-pro",
      "benchmark_name": "SWE-bench Pro",
      "setting": "default",
      "setting_name": "",
      "score": 63.2,
      "unit": "%",
      "printed": "63.2%",
      "reported_by": "anthropic",
      "label": "Sonnet 5",
      "source": "https://www.anthropic.com/news/claude-sonnet-5",
      "source_title": "Introducing Claude Sonnet 5",
      "source_date": "2026-06-30",
      "where": "benchmark table (image), row 'Agentic coding, SWE-bench Pro', column 'Sonnet 5'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "",
      "caveat": "The source doesn't say whether this is the public or the commercial set",
      "human_verified": true,
      "human_verified_on": "2026-10-17",
      "source_image": "https://frontierlog.ai/sources/849d6a908f0d46ac.png",
      "source_image_original": "https://www-cdn.anthropic.com/images/4zrzovbb/website/9941d610909f28a504e16dd5af823df172ec6035-2600x1234.png"
    },
    {
      "id": "laguna-xs-2-1:swe-bench-pro:default:1",
      "model": "laguna-xs-2-1",
      "model_name": "Laguna XS 2.1",
      "model_date": "2026-07-02",
      "event": "https://frontierlog.ai/event/2026-07-02-laguna-xs-2-1",
      "benchmark": "swe-bench-pro",
      "benchmark_name": "SWE-bench Pro",
      "setting": "default",
      "setting_name": "",
      "score": 47.6,
      "unit": "%",
      "printed": "47.6",
      "reported_by": "poolside",
      "label": "Laguna XS 2.1 (33B-A3B)",
      "source": "https://poolside.ai/blog/introducing-laguna-xs-2-1",
      "source_title": "Introducing Laguna XS 2.1",
      "source_date": "2026-07-02",
      "where": "evaluation chart 'SWE-Bench Pro' (Resolved tasks on SWE-Bench Pro), bar 'Laguna XS 2.1 (33B-A3B)', value label (animates up from 0.0 when scrolled into view; also in the bar's aria-label)",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "poolside harness on Harbor, mean pass@1 over 2 attempts, some task images patched for infrastructure reliability",
      "caveat": "The source doesn't say whether this is the public or the commercial set",
      "human_verified": true,
      "human_verified_on": "2026-10-17"
    },
    {
      "id": "gpt-5-6-sol:swe-bench-pro:default:1",
      "model": "gpt-5-6-sol",
      "model_name": "GPT-5.6 Sol",
      "model_date": "2026-07-09",
      "event": "https://frontierlog.ai/event/2026-07-09-gpt-5-6",
      "benchmark": "swe-bench-pro",
      "benchmark_name": "SWE-bench Pro",
      "setting": "default",
      "setting_name": "",
      "score": 64.6,
      "unit": "%",
      "printed": "64.6%",
      "reported_by": "openai",
      "label": "GPT‑5.6 Sol",
      "source": "https://openai.com/index/gpt-5-6",
      "source_title": "GPT-5.6: Frontier intelligence that scales with your ambition",
      "source_date": "2026-07-09",
      "where": "coding eval table (columns GPT‑5.6 Sol, Sol Ultra, Terra, Luna, GPT‑5.5, …), row 'SWE-Bench Pro', column 'GPT‑5.6 Sol'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "",
      "caveat": "The source doesn't say whether this is the public or the commercial set",
      "human_verified": true,
      "human_verified_on": "2026-10-17"
    },
    {
      "id": "inkling:swe-bench-pro:default:1",
      "model": "inkling",
      "model_name": "Inkling",
      "model_date": "2026-07-15",
      "event": "https://frontierlog.ai/event/2026-07-15-inkling",
      "benchmark": "swe-bench-pro",
      "benchmark_name": "SWE-bench Pro",
      "setting": "default",
      "setting_name": "",
      "score": 54.3,
      "unit": "%",
      "printed": "54.3%",
      "reported_by": "thinking-machines",
      "label": "Inkling effort=0.99",
      "source": "https://thinkingmachines.ai/news/introducing-inkling/",
      "source_title": "Inkling: Our Open-Weights Model",
      "source_date": "2026-07-15",
      "where": "section 'Benchmarking Inkling', main benchmark table, row 'SWEBench Pro' with subtitle 'Public' (Agentic (coding)), column 'Inkling effort=0.99' (the Inkling vs Inkling-Small table later prints the same value)",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "effort 0.99, temperature 1.0, 256K-token trajectory limit",
      "caveat": "",
      "human_verified": true,
      "human_verified_on": "2026-10-17"
    },
    {
      "id": "gemini-3-6-flash:swe-bench-pro:default:1",
      "model": "gemini-3-6-flash",
      "model_name": "Gemini 3.6 Flash",
      "model_date": "2026-07-21",
      "event": "https://frontierlog.ai/event/2026-07-21-gemini-3-6-flash-and-3-5-flash",
      "benchmark": "swe-bench-pro",
      "benchmark_name": "SWE-bench Pro",
      "setting": "default",
      "setting_name": "",
      "score": 58.7,
      "unit": "%",
      "printed": "58.7%",
      "reported_by": "google",
      "label": "Gemini 3.6 Flash",
      "source": "https://deepmind.google/models/model-cards/gemini-3-6-flash/",
      "source_title": "Gemini 3.6 Flash - Model Card",
      "source_date": "2026-07-21",
      "where": "model card, Evaluation > Results table (HTML), row 'SWE-Bench Pro (Public)', column 'Gemini 3.6 Flash'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "results as of July 2026",
      "caveat": "",
      "human_verified": true,
      "human_verified_on": "2026-10-17"
    },
    {
      "id": "claude-opus-5:swe-bench-pro:default:1",
      "model": "claude-opus-5",
      "model_name": "Claude Opus 5",
      "model_date": "2026-07-24",
      "event": "https://frontierlog.ai/event/2026-07-24-claude-opus-5",
      "benchmark": "swe-bench-pro",
      "benchmark_name": "SWE-bench Pro",
      "setting": "default",
      "setting_name": "",
      "score": 79.2,
      "unit": "%",
      "printed": "79.2",
      "reported_by": "anthropic",
      "label": "Opus 5",
      "source": "https://www.anthropic.com/claude-opus-5-system-card",
      "source_title": "Claude Opus 5 System Card",
      "source_date": "2026-07-24",
      "where": "Section 8.1 Evaluation summary, Table 8.1.A, row 'SWE-bench Pro', column 'Opus 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"
    },
    {
      "id": "qwen3-8-max:swe-bench-pro:default:1",
      "model": "qwen3-8-max",
      "model_name": "Qwen3.8-Max",
      "model_date": "2026-08-03",
      "event": "https://frontierlog.ai/event/2026-08-03-qwen3-8-max",
      "benchmark": "swe-bench-pro",
      "benchmark_name": "SWE-bench Pro",
      "setting": "default",
      "setting_name": "",
      "score": 67.7,
      "unit": "%",
      "printed": "67.7",
      "reported_by": "alibaba",
      "label": "Qwen3.8-Max",
      "source": "https://qwen.ai/blog?id=qwen3.8",
      "source_title": "Qwen3.8-Max: A New Bar for Coding and Cowork",
      "source_date": "2026-08-03",
      "where": "'Full Benchmark Table' (first table), row 'SWE-bench Pro' under 'Coding Agent', column 'Qwen3.8-Max' (read via the qwen.ai article API, qwen.ai/api/v2/article/?language=en-US&path=qwen3.8&type=qwen_ai, article 65dfab20-97fb-4932-97cd-4f7da4ab51f5)",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "Claude Code harness",
      "caveat": "Run on the lab's own corrected version of the public set",
      "human_verified": true,
      "human_verified_on": "2026-10-17"
    },
    {
      "id": "claude-fable-5-1:swe-bench-pro:default:1",
      "model": "claude-fable-5-1",
      "model_name": "Claude Fable 5.1",
      "model_date": "2026-09-01",
      "event": "https://frontierlog.ai/event/2026-09-01-claude-fable-5-1-mythos-5-1",
      "benchmark": "swe-bench-pro",
      "benchmark_name": "SWE-bench Pro",
      "setting": "default",
      "setting_name": "",
      "score": 81.2,
      "unit": "%",
      "printed": "81.2%",
      "reported_by": "anthropic",
      "label": "Fable 5.1",
      "source": "https://www.anthropic.com/claude-fable-5-1-mythos-5-1-system-card",
      "source_title": "Claude Fable 5.1 and Claude Mythos 5.1 System Card",
      "source_date": "2026-09-01",
      "where": "Section 8.2 'SWE-bench Pro, Multilingual, and Multimodal', first bullet",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "average of 5 trials; from the card's prose (its table merges Fable 5.1 and Mythos 5.1)",
      "caveat": "The source doesn't say whether this is the public or the commercial set",
      "human_verified": true,
      "human_verified_on": "2026-10-17"
    },
    {
      "id": "claude-opus-5-5:swe-bench-pro:default:1",
      "model": "claude-opus-5-5",
      "model_name": "Claude Opus 5.5",
      "model_date": "2026-09-22",
      "event": "https://frontierlog.ai/event/2026-09-22-claude-opus-5-5",
      "benchmark": "swe-bench-pro",
      "benchmark_name": "SWE-bench Pro",
      "setting": "default",
      "setting_name": "",
      "score": 89.9,
      "unit": "%",
      "printed": "89.9",
      "reported_by": "anthropic",
      "label": "Claude Opus 5.5",
      "source": "https://www.anthropic.com/claude-opus-5-5-system-card",
      "source_title": "Claude Opus 5.5 System Card",
      "source_date": "2026-09-22",
      "where": "Section 8.1 Evaluation summary, Table 8.1.A, row 'SWE-bench Pro', column 'Claude Opus 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"
    },
    {
      "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",
      "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"
    },
    {
      "id": "claude-haiku-5-5:swe-bench-pro:default:1",
      "model": "claude-haiku-5-5",
      "model_name": "Claude Haiku 5.5",
      "model_date": "2026-10-07",
      "event": "https://frontierlog.ai/event/2026-10-07-claude-haiku-5-5",
      "benchmark": "swe-bench-pro",
      "benchmark_name": "SWE-bench Pro",
      "setting": "default",
      "setting_name": "",
      "score": 64.8,
      "unit": "%",
      "printed": "64.8",
      "reported_by": "anthropic",
      "label": "Haiku 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 8.1 Evaluation summary, Table 8.1.A, row 'SWE-Bench Pro', column 'Haiku 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"
    }
  ]
}