{
  "api_version": "1",
  "updated": "2026-10-22",
  "latest_event": "2026-10-21",
  "rights": "© Frontier Log, all rights reserved.",
  "story": {
    "slug": "benchmark-scores-move-after-launch-day",
    "title": "Benchmark Scores Move After Launch Day",
    "description": "How labs re-report their older models' benchmark scores in later documents, and how far a launch-day score holds.",
    "url": "https://frontierlog.ai/stories/benchmark-scores-move-after-launch-day",
    "date": "2026-10-22",
    "image": "https://frontierlog.ai/og/story-benchmark-scores-move-after-launch-day-two-labs.png",
    "event": null,
    "tags": [],
    "images": [
      "https://frontierlog.ai/og/story-benchmark-scores-move-after-launch-day-two-labs.png",
      "https://frontierlog.ai/og/story-benchmark-scores-move-after-launch-day-labs.png",
      "https://frontierlog.ai/og/story-benchmark-scores-move-after-launch-day-launch.png",
      "https://frontierlog.ai/og/story-benchmark-scores-move-after-launch-day-biggest.png"
    ],
    "steps": [],
    "tables": [
      {
        "title": "AIME 2024, no tools",
        "unit": "%",
        "rows": [
          {
            "event": "2024-09-12-openai-o1",
            "date": "2024-09-12",
            "lab": "openai",
            "value": 44.6,
            "printed": "44.6%",
            "source": "https://openai.com/index/openai-o1-mini-advancing-cost-efficient-reasoning",
            "where": "'Optimized for STEM Reasoning' section, the 'Mathematics:' paragraph under the 'Math Performance vs Inference Cost' chart, the figure in parentheses after 'o1-preview'"
          },
          {
            "event": "2024-09-12-openai-o1",
            "date": "2024-09-12",
            "lab": "openai",
            "value": 42.0,
            "printed": "42.0",
            "source": "https://openai.com/index/o1-and-new-tools-for-developers",
            "where": "eval table (Category / Eval / o1-2024-12-17 / o1-preview), row 'Math: AIME 2024 (pass @1)', column 'o1-preview'"
          }
        ]
      },
      {
        "title": "AIME 2025, no tools",
        "unit": "%",
        "rows": [
          {
            "event": "2025-03-18-llama-nemotron",
            "date": "2025-03-18",
            "lab": "nvidia",
            "value": 58.4,
            "printed": "58.4",
            "source": "https://huggingface.co/nvidia/Llama-3_3-Nemotron-Super-49B-v1/blob/9314922284e3ef0e86ddbcad0d8b8dbd62238cfb/README.md",
            "where": "'Evaluation Results' section, 'AIME25' table, row 'Reasoning On', column 'pass@1'"
          },
          {
            "event": "2025-03-18-llama-nemotron",
            "date": "2025-03-18",
            "lab": "nvidia",
            "value": 53.3,
            "printed": "53.3",
            "source": "https://developer.nvidia.com/blog/build-more-accurate-and-efficient-ai-agents-with-the-new-nvidia-llama-nemotron-super-v1-5/",
            "where": "Figure 2 bar chart (image), value label of the 'Llama Nemotron Super 49B v1' bar (lightest, rightmost in group) in group 'AIME25'"
          },
          {
            "event": "2025-03-18-llama-nemotron",
            "date": "2025-03-18",
            "lab": "nvidia",
            "value": 58.4,
            "printed": "58.4",
            "source": "https://huggingface.co/nvidia/Llama-3_3-Nemotron-Super-49B-v1/blob/9314922284e3ef0e86ddbcad0d8b8dbd62238cfb/README.md",
            "where": "'Evaluation Results' section, 'AIME25' table, row 'Reasoning On', column 'pass@1'"
          },
          {
            "event": "2025-03-18-llama-nemotron",
            "date": "2025-03-18",
            "lab": "nvidia",
            "value": 46.0,
            "printed": "46.0",
            "source": "https://research.nvidia.com/labs/nemotron/files/NVFP4-QAD-Report.pdf",
            "where": "Table 2 (SFT-heavy models, PDF page 5), model 'Llama Nemotron Super V1', row 'BF16', column 'AIME25'"
          },
          {
            "event": "2025-03-25-gemini-2-5-pro",
            "date": "2025-03-25",
            "lab": "google",
            "value": 86.7,
            "printed": "86.7%",
            "source": "https://blog.google/innovation-and-ai/models-and-research/google-deepmind/gemini-model-thinking-updates-march-2025/",
            "where": "benchmark table image (gemini_benchmarks_cropped GIF), row 'Mathematics AIME 2025', sub-row 'single attempt (pass@1)', column 'Gemini 2.5 Pro Experimental (03-25)'"
          },
          {
            "event": "2025-03-25-gemini-2-5-pro",
            "date": "2025-03-25",
            "lab": "google",
            "value": 88.0,
            "printed": "88.0%",
            "source": "https://blog.google/innovation-and-ai/technology/developers-tools/gemini-3-developers/",
            "where": "benchmark table (image), row 'AIME 2025', sub-row 'No tools', column 'Gemini 2.5 Pro'"
          }
        ]
      },
      {
        "title": "AutomationBench",
        "unit": "%",
        "rows": [
          {
            "event": "2026-05-28-claude-opus-4-8",
            "date": "2026-05-28",
            "lab": "anthropic",
            "value": 15.5,
            "printed": "15.5",
            "source": "https://www.anthropic.com/claude-opus-4-8-system-card",
            "where": "Section 8.1, Table 8.1.A (evaluation summary), row 'Automation Bench', column 'Claude Opus 4.8'"
          },
          {
            "event": "2026-05-28-claude-opus-4-8",
            "date": "2026-05-28",
            "lab": "anthropic",
            "value": 17.0,
            "printed": "17.0%",
            "source": "https://www.anthropic.com/news/claude-opus-5",
            "where": "benchmark table (image), row 'Business workflows, AutomationBench', column 'Opus 4.8'"
          },
          {
            "event": "2026-06-09-claude-fable-5-mythos-5",
            "date": "2026-06-09",
            "lab": "anthropic",
            "value": 17.4,
            "printed": "17.4",
            "source": "https://www.anthropic.com/claude-fable-5-mythos-5-system-card",
            "where": "Section 8.1, Table 8.1.A, row 'AutomationBench', column 'Fable 5'"
          },
          {
            "event": "2026-06-09-claude-fable-5-mythos-5",
            "date": "2026-06-09",
            "lab": "anthropic",
            "value": 17.1,
            "printed": "17.1%",
            "source": "https://www.anthropic.com/claude-fable-and-mythos-5-1",
            "where": "benchmark table (HTML), row 'Business workflows, AutomationBench', column 'Fable 5'"
          },
          {
            "event": "2026-06-30-claude-sonnet-5",
            "date": "2026-06-30",
            "lab": "anthropic",
            "value": 13.5,
            "printed": "13.5",
            "source": "https://www.anthropic.com/claude-sonnet-5-system-card",
            "where": "Section 8.1, Table 8.1.A, row 'AutomationBench', column 'Claude Sonnet 5'"
          },
          {
            "event": "2026-06-30-claude-sonnet-5",
            "date": "2026-06-30",
            "lab": "anthropic",
            "value": 10.7,
            "printed": "10.7",
            "source": "https://www.anthropic.com/claude-sonnet-5-5-system-card",
            "where": "Section 8.1, Table 8.1.A, row 'AutomationBench', column 'Claude Sonnet 5'"
          },
          {
            "event": "2026-07-24-claude-opus-5",
            "date": "2026-07-24",
            "lab": "anthropic",
            "value": 26.0,
            "printed": "26.0%",
            "source": "https://www.anthropic.com/news/claude-opus-5",
            "where": "benchmark table (image), row 'Business workflows, AutomationBench', column 'Opus 5'"
          },
          {
            "event": "2026-07-24-claude-opus-5",
            "date": "2026-07-24",
            "lab": "anthropic",
            "value": 26.9,
            "printed": "26.9%",
            "source": "https://www.anthropic.com/claude-fable-and-mythos-5-1",
            "where": "benchmark table (HTML), row 'Business workflows, AutomationBench', column 'Opus 5'"
          },
          {
            "event": "2026-08-13-deepseek-v4-pro-ga",
            "date": "2026-08-13",
            "lab": "deepseek",
            "value": 31.8,
            "printed": "31.8",
            "source": "https://api-docs.deepseek.com/news/news260813/",
            "where": "table image, row 'AutomationBench (Public)', column 'DeepSeek-V4-Pro-0813'"
          },
          {
            "event": "2026-08-13-deepseek-v4-pro-ga",
            "date": "2026-08-13",
            "lab": "deepseek",
            "value": 43.2,
            "printed": "43.2",
            "source": "https://api-docs.deepseek.com/news/news260910/",
            "where": "later report: DeepSeek-V4.1-Flash post, benchmark table image (second image), row 'Automation-Bench', column 'DeepSeek V4-Pro 0813'"
          },
          {
            "event": "2026-09-03-gpt-6-astra",
            "date": "2026-09-03",
            "lab": "openai",
            "value": 41.4,
            "printed": "41.4%",
            "source": "https://openai.com/index/gpt-6-astra",
            "where": "Professional table, row 'AutomationBench', column 'GPT‑6 Astra'"
          },
          {
            "event": "2026-09-03-gpt-6-astra",
            "date": "2026-09-03",
            "lab": "openai",
            "value": 30.3,
            "printed": "30.3%",
            "source": "https://openai.com/index/introducing-gpt-6-sol-and-luna",
            "where": "AutomationBench table (Model (and effort) / Score / Cost per task), row 'GPT-6 Astra (low)', column 'Score'"
          },
          {
            "event": "2026-09-22-claude-opus-5-5",
            "date": "2026-09-22",
            "lab": "anthropic",
            "value": 40.0,
            "printed": "40.0%",
            "source": "https://www.anthropic.com/claude-opus-5-5",
            "where": "benchmark table, row 'AutomationBench', column 'Opus 5.5'"
          },
          {
            "event": "2026-09-22-claude-opus-5-5",
            "date": "2026-09-22",
            "lab": "anthropic",
            "value": 42.5,
            "printed": "42.5",
            "source": "https://www.anthropic.com/claude-sonnet-5-5-system-card",
            "where": "Section 8.1, Table 8.1.A, row 'AutomationBench', column 'Claude Opus 5.5'"
          }
        ]
      },
      {
        "title": "BrowseComp",
        "unit": "%",
        "rows": [
          {
            "event": "2025-09-30-glm-4-6",
            "date": "2025-09-30",
            "lab": "zai",
            "value": 45.1,
            "printed": "45.1",
            "source": "https://z.ai/blog/glm-4.6",
            "where": "benchmark chart image (eight panels), panel 'BrowseComp', first bar 'GLM-4.6', value label"
          },
          {
            "event": "2025-09-30-glm-4-6",
            "date": "2025-09-30",
            "lab": "zai",
            "value": 57.5,
            "printed": "57.5",
            "source": "https://z.ai/blog/glm-4.7",
            "where": "benchmark table, row 'BrowseComp w/ Context Manage', column 'GLM-4.6'"
          },
          {
            "event": "2025-10-27-minimax-m2",
            "date": "2025-10-27",
            "lab": "minimax",
            "value": 44.0,
            "printed": "44.0",
            "source": "https://www.minimax.io/blog/minimax-m2-en-1748600000",
            "where": "first benchmark chart image (wide, eight panels), panel 'BrowseComp', first bar 'MiniMax-M2' (red), value label"
          },
          {
            "event": "2025-10-27-minimax-m2",
            "date": "2025-10-27",
            "lab": "minimax",
            "value": 56.9,
            "printed": "56.9",
            "source": "https://www.minimax.io/news/minimax-m21",
            "where": "'Benchmarks' section, last (long-horizon tool use and intelligence) table image, row 'BrowseComp (context management)', column 'MiniMax-M2'"
          },
          {
            "event": "2026-02-17-claude-sonnet-4-6",
            "date": "2026-02-17",
            "lab": "anthropic",
            "value": 74.7,
            "printed": "74.7%",
            "source": "https://www.anthropic.com/news/claude-sonnet-4-6",
            "where": "benchmark table (image), row 'Agentic search, BrowseComp', column 'Sonnet 4.6'"
          },
          {
            "event": "2026-02-17-claude-sonnet-4-6",
            "date": "2026-02-17",
            "lab": "anthropic",
            "value": 76.2,
            "printed": "76.2",
            "source": "https://www.anthropic.com/claude-sonnet-5-system-card",
            "where": "Section 8.1 evaluation summary, Table 8.1.A (page 115), row 'BrowseComp', column 'Claude Sonnet 4.6'"
          },
          {
            "event": "2026-03-11-nemotron-3-super",
            "date": "2026-03-11",
            "lab": "nvidia",
            "value": 31.28,
            "printed": "31.28",
            "source": "https://research.nvidia.com/labs/nemotron/files/NVIDIA-Nemotron-3-Super-Technical-Report.pdf",
            "where": "Table 5 (PDF page 31), row 'BrowseComp with Search', column 'N-3-Super'"
          },
          {
            "event": "2026-03-11-nemotron-3-super",
            "date": "2026-03-11",
            "lab": "nvidia",
            "value": 22.77,
            "printed": "22.77",
            "source": "https://research.nvidia.com/labs/nemotron/files/NVIDIA-Nemotron-3.5-Bolt-74B-A9B-24GB-Technical-Report.pdf",
            "where": "Table 7 (PDF page 16), row 'BrowseComp', column 'Nemotron 3 Super 120B-A12B-BF16' (last column)"
          },
          {
            "event": "2026-04-07-claude-mythos-preview",
            "date": "2026-04-07",
            "lab": "anthropic",
            "value": 86.9,
            "printed": "86.9%",
            "source": "https://www.anthropic.com/news/claude-opus-4-7",
            "where": "benchmark table (image), row 'Agentic search, BrowseComp', column 'Mythos Preview'"
          },
          {
            "event": "2026-04-07-claude-mythos-preview",
            "date": "2026-04-07",
            "lab": "anthropic",
            "value": 87.9,
            "printed": "87.9",
            "source": "https://www.anthropic.com/claude-fable-5-mythos-5-system-card",
            "where": "Section 8.1 evaluation summary table (page 251), row 'BrowseComp', column 'Mythos Preview'"
          },
          {
            "event": "2026-04-16-claude-opus-4-7",
            "date": "2026-04-16",
            "lab": "anthropic",
            "value": 79.3,
            "printed": "79.3%",
            "source": "https://www.anthropic.com/news/claude-opus-4-7",
            "where": "benchmark table (image), row 'Agentic search, BrowseComp', column 'Opus 4.7'"
          },
          {
            "event": "2026-04-16-claude-opus-4-7",
            "date": "2026-04-16",
            "lab": "anthropic",
            "value": 79.8,
            "printed": "79.8",
            "source": "https://www.anthropic.com/claude-opus-4-8-system-card",
            "where": "Section 8.1 evaluation summary table (page 194), row 'BrowseComp', column 'Claude Opus 4.7'"
          },
          {
            "event": "2026-06-04-nemotron-3-ultra",
            "date": "2026-06-04",
            "lab": "nvidia",
            "value": 44.4,
            "printed": "44.4",
            "source": "https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4/blob/02c7d9e60f328a4fb9d990d6b613057b19ab8a8f/README.md",
            "where": "benchmark table, row 'BrowseComp', column 'Nemotron 3 Ultra BF16'"
          },
          {
            "event": "2026-06-04-nemotron-3-ultra",
            "date": "2026-06-04",
            "lab": "nvidia",
            "value": 61.06,
            "printed": "61.06",
            "source": "https://research.nvidia.com/labs/nemotron/files/NVIDIA-Nemotron-3.5-Super-and-Lightning-Technical-Report.pdf",
            "where": "Table 10 (PDF page 23), row 'BrowseComp (OpenCode)', column 'N3-Ultra 550B-A55B' (second column)"
          },
          {
            "event": "2026-08-11-nemotron-3-5-lightning",
            "date": "2026-08-11",
            "lab": "nvidia",
            "value": 36.97,
            "printed": "36.97",
            "source": "https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-NVFP4/blob/b14872a58afd94e2fbbc1e5a1fe4ebb7dfdc5fdd/README.md",
            "where": "'Reasoning Benchmark Evaluations' table, row 'BrowseComp', column 'Nemotron-3.5-Lightning-30B-A3B-BF16'"
          },
          {
            "event": "2026-08-11-nemotron-3-5-lightning",
            "date": "2026-08-11",
            "lab": "nvidia",
            "value": 39.5,
            "printed": "39.50",
            "source": "https://research.nvidia.com/labs/nemotron/files/NVIDIA-Nemotron-3.5-Super-and-Lightning-Technical-Report.pdf",
            "where": "Table 22 (PDF page 42), row 'BrowseComp', column 'BF16 acc.'"
          }
        ]
      },
      {
        "title": "DeepSWE v1.1",
        "unit": "%",
        "rows": [
          {
            "event": "2026-07-09-muse-spark-1-1",
            "date": "2026-07-09",
            "lab": "meta",
            "value": 53.3,
            "printed": "53.3",
            "source": "https://ai.meta.com/blog/introducing-muse-spark-meta-model-api/",
            "where": "benchmark table (image), row 'DeepSWE 1.1', column 'Muse Spark 1.1'"
          },
          {
            "event": "2026-07-09-muse-spark-1-1",
            "date": "2026-07-09",
            "lab": "meta",
            "value": 53.0,
            "printed": "53.0%",
            "source": "https://research.meta.ai/blog/introducing-muse-code-and-muse-spark-1-2",
            "where": "bar chart 'DeepSWE 1.1' (image, Coding tab), value label above the bar 'Muse Spark 1.1, mini-swe-agent'"
          },
          {
            "event": "2026-07-21-gemini-3-6-flash-and-3-5-flash",
            "date": "2026-07-21",
            "lab": "google",
            "value": 49.0,
            "printed": "49%",
            "source": "https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-6-flash-3-5-flash-lite-3-5-flash-cyber/",
            "where": "chart 'Gemini 3.6 Flash: Outperforms prior generations across agentic benchmarks' (image), panel 'Long-horizon software engineering, DeepSWE v1.1', bar '3.6 Flash' value label"
          },
          {
            "event": "2026-07-21-gemini-3-6-flash-and-3-5-flash",
            "date": "2026-07-21",
            "lab": "google",
            "value": 48.6,
            "printed": "48.6%",
            "source": "https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-3-7-flash/",
            "where": "benchmark table (image), row 'DeepSWE v1.1', column 'Gemini 3.6 Flash'"
          },
          {
            "event": "2026-08-05-muse-spark-1-2",
            "date": "2026-08-05",
            "lab": "meta",
            "value": 59.3,
            "printed": "59.3%",
            "source": "https://research.meta.ai/blog/introducing-muse-code-and-muse-spark-1-2",
            "where": "bar chart 'DeepSWE 1.1' (image, Coding tab), value label above the bar 'Muse Spark 1.2, Muse Code'"
          },
          {
            "event": "2026-08-05-muse-spark-1-2",
            "date": "2026-08-05",
            "lab": "meta",
            "value": 55.0,
            "printed": "55.0",
            "source": "https://research.meta.ai/blog/introducing-muse-spark-1-3",
            "where": "Muse Spark 1.3 scorecard (image), row 'DeepSWE v1.1', column 'Muse Spark 1.2 (xhigh)'"
          }
        ]
      },
      {
        "title": "FrontierCode 1.1, main set",
        "unit": "%",
        "rows": [
          {
            "event": "2026-07-24-claude-opus-5",
            "date": "2026-07-24",
            "lab": "anthropic",
            "value": 53.4,
            "printed": "53.4%",
            "source": "https://www.anthropic.com/news/claude-opus-5",
            "where": "benchmark table (image), row 'Agentic coding, FrontierCode v1.1, Main', column 'Opus 5'"
          },
          {
            "event": "2026-07-24-claude-opus-5",
            "date": "2026-07-24",
            "lab": "anthropic",
            "value": 48.0,
            "printed": "48.0%",
            "source": "https://www.anthropic.com/claude-opus-5-5",
            "where": "benchmark table (HTML), row 'Agentic coding, FrontierCode v1.1 (Main)', column 'Opus 5'"
          },
          {
            "event": "2026-09-01-claude-fable-5-1-mythos-5-1",
            "date": "2026-09-01",
            "lab": "anthropic",
            "value": 50.9,
            "printed": "50.9%",
            "source": "https://www.anthropic.com/claude-fable-5-1-mythos-5-1-system-card",
            "where": "Section 8.4 'FrontierCode', paragraph after the task description, sentence on FrontierCode 1.1 (Main)"
          },
          {
            "event": "2026-09-01-claude-fable-5-1-mythos-5-1",
            "date": "2026-09-01",
            "lab": "anthropic",
            "value": 50.3,
            "printed": "50.3%",
            "source": "https://www.anthropic.com/claude-opus-5-5",
            "where": "benchmark table (HTML), row 'Agentic coding, FrontierCode v1.1 (Main)', column 'Fable 5.1'"
          },
          {
            "event": "2026-09-01-claude-fable-5-1-mythos-5-1",
            "date": "2026-09-01",
            "lab": "anthropic",
            "value": 50.9,
            "printed": "50.9%",
            "source": "https://www.anthropic.com/claude-fable-5-1-mythos-5-1-system-card",
            "where": "Section 8.4 'FrontierCode', paragraph after the task description, sentence on FrontierCode 1.1 (Main)"
          },
          {
            "event": "2026-09-01-claude-fable-5-1-mythos-5-1",
            "date": "2026-09-01",
            "lab": "anthropic",
            "value": 52.8,
            "printed": "52.8%",
            "source": "https://www.anthropic.com/claude-opus-5-5-system-card",
            "where": "Section 8.4 'FrontierCode', paragraph ranking Opus 5.5, the Claude Fable 5.1 figure in the FrontierCode (Main) sentence"
          }
        ]
      },
      {
        "title": "GPQA Diamond",
        "unit": "%",
        "rows": [
          {
            "event": "2024-05-13-gpt-4o",
            "date": "2024-05-13",
            "lab": "openai",
            "value": 53.6,
            "printed": "53.6",
            "source": "https://openai.com/index/hello-gpt-4o",
            "where": "text evaluation bar chart (SVG), panel 'GPQA (%)', bar 'GPT-4o'"
          },
          {
            "event": "2024-05-13-gpt-4o",
            "date": "2024-05-13",
            "lab": "openai",
            "value": 50.6,
            "printed": "50.6",
            "source": "https://openai.com/index/learning-to-reason-with-llms",
            "where": "eval table (GPT-4o, o1-preview, o1), row 'GPQA Diamond', sub-row 'pass@1', column 'GPT-4o'"
          },
          {
            "event": "2024-07-23-llama-3-1-405b",
            "date": "2024-07-23",
            "lab": "meta",
            "value": 51.1,
            "printed": "51.1",
            "source": "https://ai.meta.com/blog/meta-llama-3-1/",
            "where": "benchmark table (image), row 'GPQA (0-shot, CoT)', column 'Llama 3.1 405B'"
          },
          {
            "event": "2024-07-23-llama-3-1-405b",
            "date": "2024-07-23",
            "lab": "meta",
            "value": 49.0,
            "printed": "49.0",
            "source": "https://ai.meta.com/blog/llama-4-multimodal-intelligence/",
            "where": "'Llama 4 Scout instruction-tuned benchmarks' table (image), row 'GPQA Diamond', column 'Llama 3.1 405B'"
          },
          {
            "event": "2024-12-17-o1-in-api",
            "date": "2024-12-17",
            "lab": "openai",
            "value": 75.7,
            "printed": "75.7",
            "source": "https://openai.com/index/o1-and-new-tools-for-developers",
            "where": "eval table (Category / Eval / o1-2024-12-17 / o1-preview), row 'General: GPQA diamond', column 'o1-2024-12-17'"
          },
          {
            "event": "2024-12-17-o1-in-api",
            "date": "2024-12-17",
            "lab": "openai",
            "value": 78.0,
            "printed": "78.0",
            "source": "https://openai.com/index/introducing-o3-and-o4-mini",
            "where": "'What's changed' section, bar chart 'GPQA Diamond / PhD-Level Science Questions', 1st bar 'o1'"
          },
          {
            "event": "2025-01-31-openai-o3-mini",
            "date": "2025-01-31",
            "lab": "openai",
            "value": 79.7,
            "printed": "79.7",
            "source": "https://openai.com/index/openai-o3-mini",
            "where": "image GPQA.png (bar chart, accuracy by OpenAI model), value label over the 'o3-mini (high)' bar"
          },
          {
            "event": "2025-01-31-openai-o3-mini",
            "date": "2025-01-31",
            "lab": "openai",
            "value": 77.2,
            "printed": "77.2%",
            "source": "https://openai.com/index/gpt-4-1",
            "where": "academic eval table (columns GPT-4.1 ... GPT-4.5), row 'GPQA Diamond1', column 'OpenAI o3-mini(high)'"
          },
          {
            "event": "2025-01-31-openai-o3-mini",
            "date": "2025-01-31",
            "lab": "openai",
            "value": 79.7,
            "printed": "79.7",
            "source": "https://openai.com/index/openai-o3-mini",
            "where": "image GPQA.png (bar chart, accuracy by OpenAI model), value label over the 'o3-mini (high)' bar"
          },
          {
            "event": "2025-01-31-openai-o3-mini",
            "date": "2025-01-31",
            "lab": "openai",
            "value": 77.0,
            "printed": "77.0",
            "source": "https://openai.com/index/introducing-o3-and-o4-mini",
            "where": "'What's changed' section, bar chart 'GPQA Diamond / PhD-Level Science Questions', 2nd bar 'o3-mini' (page screenshot (job 20879 capture), part 02)"
          },
          {
            "event": "2025-02-27-gpt-4-5",
            "date": "2025-02-27",
            "lab": "openai",
            "value": 71.4,
            "printed": "71.4%",
            "source": "https://openai.com/index/introducing-gpt-4-5",
            "where": "eval table, row 'GPQA (science)', column 'GPT‑4.5'"
          },
          {
            "event": "2025-02-27-gpt-4-5",
            "date": "2025-02-27",
            "lab": "openai",
            "value": 69.5,
            "printed": "69.5%",
            "source": "https://openai.com/index/gpt-4-1",
            "where": "academic eval table (columns GPT-4.1 ... GPT-4.5), row 'GPQA Diamond1', column 'GPT-4.5' (last column)"
          },
          {
            "event": "2025-03-25-gemini-2-5-pro",
            "date": "2025-03-25",
            "lab": "google",
            "value": 84.0,
            "printed": "84.0%",
            "source": "https://blog.google/innovation-and-ai/models-and-research/google-deepmind/gemini-model-thinking-updates-march-2025/",
            "where": "benchmark table (image), row 'Science, GPQA diamond', 'single attempt (pass@1)', column 'Gemini 2.5 Pro Experimental (03-25)'"
          },
          {
            "event": "2025-03-25-gemini-2-5-pro",
            "date": "2025-03-25",
            "lab": "google",
            "value": 86.4,
            "printed": "86.4%",
            "source": "https://blog.google/innovation-and-ai/technology/developers-tools/gemini-3-developers/",
            "where": "benchmark table (image), row 'GPQA Diamond', 'No tools', column 'Gemini 2.5 Pro'"
          },
          {
            "event": "2025-05-07-mistral-medium-3",
            "date": "2025-05-07",
            "lab": "mistral",
            "value": 57.1,
            "printed": "57.1%",
            "source": "https://mistral.ai/news/mistral-medium-3/",
            "where": "benchmark table image ('Table Medium 6'), section KNOWLEDGE, row 'GPQA Diamond 5-shot CoT', column 'Mistral Medium 3'"
          },
          {
            "event": "2025-05-07-mistral-medium-3",
            "date": "2025-05-07",
            "lab": "mistral",
            "value": 59.6,
            "printed": "59.6%",
            "source": "https://mistral.ai/news/magistral/",
            "where": "later report: Magistral post, bar chart image ('Plot Magistral'), group 'GPQA Diamond', first bar (legend 'Mistral-Medium 3'), value label"
          },
          {
            "event": "2025-07-28-glm-4-5",
            "date": "2025-07-28",
            "lab": "zai",
            "value": 79.1,
            "printed": "79.1",
            "source": "https://z.ai/blog/glm-4.5",
            "where": "'Reasoning' section table, row 'GPQA', column 'GLM-4.5'"
          },
          {
            "event": "2025-07-28-glm-4-5",
            "date": "2025-07-28",
            "lab": "zai",
            "value": 79.9,
            "printed": "79.9",
            "source": "https://z.ai/blog/glm-4.6",
            "where": "GLM-4.6 post, benchmark chart image (eight panels), panel 'GPQA', second bar 'GLM-4.5', value label"
          },
          {
            "event": "2025-03-18-llama-nemotron",
            "date": "2025-03-18",
            "lab": "nvidia",
            "value": 66.7,
            "printed": "66.7",
            "source": "https://developer.nvidia.com/blog/build-more-accurate-and-efficient-ai-agents-with-the-new-nvidia-llama-nemotron-super-v1-5/",
            "where": "Figure 2 bar chart (image), value label of the 'Llama Nemotron Super 49B v1' bar (lightest, rightmost in group) in group 'GPQA Diamond'"
          },
          {
            "event": "2025-03-18-llama-nemotron",
            "date": "2025-03-18",
            "lab": "nvidia",
            "value": 66.5,
            "printed": "66.5",
            "source": "https://research.nvidia.com/labs/nemotron/files/NVFP4-QAD-Report.pdf",
            "where": "Table 2 (SFT-heavy models, PDF page 5), model 'Llama Nemotron Super V1', row 'BF16', column 'GPQA-D'"
          },
          {
            "event": "2025-08-05-claude-opus-4-1",
            "date": "2025-08-05",
            "lab": "anthropic",
            "value": 80.9,
            "printed": "80.9%",
            "source": "https://www.anthropic.com/news/claude-opus-4-1",
            "where": "benchmark table (image), row 'Graduate-level reasoning, GPQA Diamond', column 'Claude Opus 4.1'"
          },
          {
            "event": "2025-08-05-claude-opus-4-1",
            "date": "2025-08-05",
            "lab": "anthropic",
            "value": 81.0,
            "printed": "81.0%",
            "source": "https://www.anthropic.com/news/claude-opus-4-5",
            "where": "benchmark table (image), row 'Graduate-level reasoning, GPQA Diamond', column 'Opus 4.1'"
          },
          {
            "event": "2025-12-15-nemotron-3",
            "date": "2025-12-15",
            "lab": "nvidia",
            "value": 73.04,
            "printed": "73.04",
            "source": "https://research.nvidia.com/labs/nemotron/files/NVIDIA-Nemotron-3-Nano-Technical-Report.pdf",
            "where": "Table 3 (PDF page 22), row 'GPQA (no tools)', column 'N-3-Nano'"
          },
          {
            "event": "2025-12-15-nemotron-3",
            "date": "2025-12-15",
            "lab": "nvidia",
            "value": 74.05,
            "printed": "74.05",
            "source": "https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16/blob/942b56629bd9b52c6813bb5f68bb5d10ba8cca68/README.md",
            "where": "'Reasoning Benchmark Evaluations' table, row 'GPQA Diamond (no tools)', column 'Nemotron 3 Nano*'"
          },
          {
            "event": "2026-03-05-gpt-5-4",
            "date": "2026-03-05",
            "lab": "openai",
            "value": 92.8,
            "printed": "92.8%",
            "source": "https://openai.com/index/introducing-gpt-5-4",
            "where": "academic table, row 'GPQA Diamond', column 'GPT‑5.4'"
          },
          {
            "event": "2026-03-05-gpt-5-4",
            "date": "2026-03-05",
            "lab": "openai",
            "value": 93.0,
            "printed": "93.0%",
            "source": "https://openai.com/index/introducing-gpt-5-4-mini-and-nano",
            "where": "academic table (columns GPT-5.4 (xhigh), GPT-5.4 mini (xhigh), GPT-5.4 nano (xhigh), GPT-5 mini), row 'GPQA Diamond', column 'GPT-5.4 (xhigh)'"
          },
          {
            "event": "2026-03-11-nemotron-3-super",
            "date": "2026-03-11",
            "lab": "nvidia",
            "value": 79.23,
            "printed": "79.23",
            "source": "https://research.nvidia.com/labs/nemotron/files/NVIDIA-Nemotron-3-Super-Technical-Report.pdf",
            "where": "Table 5 (PDF page 31), row 'GPQA (no tools)', column 'N-3-Super'"
          },
          {
            "event": "2026-03-11-nemotron-3-super",
            "date": "2026-03-11",
            "lab": "nvidia",
            "value": 78.6,
            "printed": "78.60",
            "source": "https://research.nvidia.com/labs/nemotron/files/NVIDIA-Nemotron-3.5-Bolt-74B-A9B-24GB-Technical-Report.pdf",
            "where": "Table 7 (PDF page 16), row 'GPQA Diamond', column 'Nemotron 3 Super 120B-A12B-BF16' (last column)"
          },
          {
            "event": "2026-08-11-nemotron-3-5-lightning",
            "date": "2026-08-11",
            "lab": "nvidia",
            "value": 75.44,
            "printed": "75.44",
            "source": "https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-NVFP4/blob/b14872a58afd94e2fbbc1e5a1fe4ebb7dfdc5fdd/README.md",
            "where": "'Reasoning Benchmark Evaluations' table, row 'GPQA Diamond (no tools)', column 'Nemotron-3.5-Lightning-30B-A3B-BF16'"
          },
          {
            "event": "2026-08-11-nemotron-3-5-lightning",
            "date": "2026-08-11",
            "lab": "nvidia",
            "value": 76.45,
            "printed": "76.45",
            "source": "https://research.nvidia.com/labs/nemotron/files/NVIDIA-Nemotron-3.5-Super-and-Lightning-Technical-Report.pdf",
            "where": "Table 22 (PDF page 42), row 'GPQA Diamond (no tools)', column 'BF16 acc.'"
          }
        ]
      },
      {
        "title": "HLE, no tools",
        "unit": "%",
        "rows": [
          {
            "event": "2025-01-31-openai-o3-mini",
            "date": "2025-01-31",
            "lab": "openai",
            "value": 13.0,
            "printed": "13.0",
            "source": "https://openai.com/index/introducing-deep-research",
            "where": "Humanity's Last Exam table (Model / Accuracy (%)), row 'OpenAI o3-mini (high)*'"
          },
          {
            "event": "2025-01-31-openai-o3-mini",
            "date": "2025-01-31",
            "lab": "openai",
            "value": 13.4,
            "printed": "13.40",
            "source": "https://openai.com/index/introducing-o3-and-o4-mini",
            "where": "'What's changed' section, bar chart 'Humanity's Last Exam', 2nd bar 'o3-mini' (page screenshot (job 20879 capture), part 02)"
          },
          {
            "event": "2025-03-25-gemini-2-5-pro",
            "date": "2025-03-25",
            "lab": "google",
            "value": 18.8,
            "printed": "18.8%",
            "source": "https://blog.google/innovation-and-ai/models-and-research/google-deepmind/gemini-model-thinking-updates-march-2025/",
            "where": "benchmark table (image), row 'Reasoning & knowledge, Humanity's Last Exam (no tools)', column 'Gemini 2.5 Pro Experimental (03-25)'"
          },
          {
            "event": "2025-03-25-gemini-2-5-pro",
            "date": "2025-03-25",
            "lab": "google",
            "value": 21.6,
            "printed": "21.6%",
            "source": "https://blog.google/innovation-and-ai/technology/developers-tools/gemini-3-developers/",
            "where": "benchmark table (image), row 'Humanity's Last Exam', 'No tools', column 'Gemini 2.5 Pro'"
          },
          {
            "event": "2025-04-16-o3-and-o4-mini",
            "date": "2025-04-16",
            "lab": "openai",
            "value": 20.32,
            "printed": "20.32",
            "source": "https://openai.com/index/introducing-o3-and-o4-mini",
            "where": "'What's changed' section, bar chart 'Humanity's Last Exam', 3rd bar 'o3 (no tools)' (page screenshot (job 20879 capture), part 02)"
          },
          {
            "event": "2025-04-16-o3-and-o4-mini",
            "date": "2025-04-16",
            "lab": "openai",
            "value": 14.7,
            "printed": "14.7%",
            "source": "https://openai.com/index/introducing-gpt-5",
            "where": "'Evaluations' section, bar chart 'Humanity's Last Exam (Full Set)*', 9th bar 'OpenAI o3 (no tools)' (page screenshot (job 20879 capture), part 05)"
          },
          {
            "event": "2025-04-16-o3-and-o4-mini",
            "date": "2025-04-16",
            "lab": "openai",
            "value": 20.32,
            "printed": "20.32",
            "source": "https://openai.com/index/introducing-o3-and-o4-mini",
            "where": "'What's changed' section, bar chart 'Humanity's Last Exam', 3rd bar 'o3 (no tools)' (page screenshot (job 20879 capture), part 02)"
          },
          {
            "event": "2025-04-16-o3-and-o4-mini",
            "date": "2025-04-16",
            "lab": "openai",
            "value": 20.2,
            "printed": "20.2%",
            "source": "https://openai.com/index/introducing-gpt-5-for-developers",
            "where": "intelligence table (columns GPT-5 (high) ... GPT-4.1 nano), row 'HLE[1] (no tools)', column 'OpenAI o3 (high)'"
          },
          {
            "event": "2025-12-15-nemotron-3",
            "date": "2025-12-15",
            "lab": "nvidia",
            "value": 10.2,
            "printed": "10.2",
            "source": "https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-FP8/blob/fa5699ff0bb9da723217af52488c0f9653512ffb/README.md",
            "where": "'Reasoning Benchmark Evaluations' table, row 'HLE (no tools)', column 'NVIDIA-Nemotron-3-Nano-30B-A3B-BF16'"
          },
          {
            "event": "2025-12-15-nemotron-3",
            "date": "2025-12-15",
            "lab": "nvidia",
            "value": 10.89,
            "printed": "10.89",
            "source": "https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16/blob/942b56629bd9b52c6813bb5f68bb5d10ba8cca68/README.md",
            "where": "'Reasoning Benchmark Evaluations' table, row 'HLE (text-only, no tools)', column 'Nemotron 3 Nano*'"
          },
          {
            "event": "2026-02-17-claude-sonnet-4-6",
            "date": "2026-02-17",
            "lab": "anthropic",
            "value": 33.2,
            "printed": "33.2%",
            "source": "https://www.anthropic.com/news/claude-sonnet-4-6",
            "where": "benchmark table (image), row 'Multidisciplinary reasoning, Humanity's Last Exam (HLE)', column 'Sonnet 4.6', 'without tools'"
          },
          {
            "event": "2026-02-17-claude-sonnet-4-6",
            "date": "2026-02-17",
            "lab": "anthropic",
            "value": 34.6,
            "printed": "34.6%",
            "source": "https://www.anthropic.com/news/claude-sonnet-5",
            "where": "benchmark table (image), row 'Multidisciplinary reasoning, Humanity's Last Exam', column 'Sonnet 4.6', 'no tools'"
          },
          {
            "event": "2026-03-11-nemotron-3-super",
            "date": "2026-03-11",
            "lab": "nvidia",
            "value": 18.26,
            "printed": "18.26",
            "source": "https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-FP8/blob/1408abd63ba6ba1c37a78b94d332dac35da45370/README.md",
            "where": "benchmark table under the accuracy chart, row 'HLE (no tools)', column 'Nemotron-3-Super'"
          },
          {
            "event": "2026-03-11-nemotron-3-super",
            "date": "2026-03-11",
            "lab": "nvidia",
            "value": 20.3,
            "printed": "20.30",
            "source": "https://research.nvidia.com/labs/nemotron/files/NVIDIA-Nemotron-3.5-Bolt-74B-A9B-24GB-Technical-Report.pdf",
            "where": "Table 7 (PDF page 16), row 'HLE (no tools)', column 'Nemotron 3 Super 120B-A12B-BF16' (last column)"
          },
          {
            "event": "2026-06-04-nemotron-3-ultra",
            "date": "2026-06-04",
            "lab": "nvidia",
            "value": 26.7,
            "printed": "26.7",
            "source": "https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4/blob/02c7d9e60f328a4fb9d990d6b613057b19ab8a8f/README.md",
            "where": "benchmark table, row 'HLE (no tools)', column 'Nemotron 3 Ultra BF16'"
          },
          {
            "event": "2026-06-04-nemotron-3-ultra",
            "date": "2026-06-04",
            "lab": "nvidia",
            "value": 29.24,
            "printed": "29.24",
            "source": "https://research.nvidia.com/labs/nemotron/files/NVIDIA-Nemotron-3.5-Super-and-Lightning-Technical-Report.pdf",
            "where": "Table 10 (PDF page 23), row 'HLE (no tools, text)', column 'N3-Ultra 550B-A55B' (second column)"
          },
          {
            "event": "2026-06-30-claude-sonnet-5",
            "date": "2026-06-30",
            "lab": "anthropic",
            "value": 43.2,
            "printed": "43.2%",
            "source": "https://www.anthropic.com/news/claude-sonnet-5",
            "where": "benchmark table (image), row 'Multidisciplinary reasoning, Humanity's Last Exam', column 'Sonnet 5', 'no tools'"
          },
          {
            "event": "2026-06-30-claude-sonnet-5",
            "date": "2026-06-30",
            "lab": "anthropic",
            "value": 43.1,
            "printed": "43.1",
            "source": "https://www.anthropic.com/claude-sonnet-5-5-system-card",
            "where": "Section 8.1 evaluation summary table, row 'Humanity's Last Exam, No tools', column 'Claude Sonnet 5'"
          },
          {
            "event": "2026-06-09-claude-fable-5-mythos-5",
            "date": "2026-06-09",
            "lab": "anthropic",
            "value": 56.5,
            "printed": "56.5%",
            "source": "https://www.anthropic.com/news/claude-opus-5",
            "where": "benchmark table (image), row 'Multidisciplinary reasoning, Humanity's Last Exam', column 'Fable 5', 'no tools'"
          },
          {
            "event": "2026-06-09-claude-fable-5-mythos-5",
            "date": "2026-06-09",
            "lab": "anthropic",
            "value": 57.8,
            "printed": "57.8%",
            "source": "https://www.anthropic.com/claude-fable-and-mythos-5-1",
            "where": "benchmark table (HTML), row 'Multidisciplinary reasoning, Humanity's Last Exam', column 'Fable 5', 'no tools'"
          },
          {
            "event": "2026-07-24-claude-opus-5",
            "date": "2026-07-24",
            "lab": "anthropic",
            "value": 56.3,
            "printed": "56.3%",
            "source": "https://www.anthropic.com/news/claude-opus-5",
            "where": "benchmark table (image), row 'Multidisciplinary reasoning, Humanity's Last Exam', column 'Opus 5', 'no tools'"
          },
          {
            "event": "2026-07-24-claude-opus-5",
            "date": "2026-07-24",
            "lab": "anthropic",
            "value": 56.6,
            "printed": "56.6",
            "source": "https://www.anthropic.com/claude-opus-5-5-system-card",
            "where": "Section 8.1 evaluation summary table, row 'Humanity's Last Exam, No tools', column 'Claude Opus 5'"
          },
          {
            "event": "2026-08-11-nemotron-3-5-lightning",
            "date": "2026-08-11",
            "lab": "nvidia",
            "value": 11.72,
            "printed": "11.72",
            "source": "https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-NVFP4/blob/b14872a58afd94e2fbbc1e5a1fe4ebb7dfdc5fdd/README.md",
            "where": "'Reasoning Benchmark Evaluations' table, row 'HLE (text-only, no tools)', column 'Nemotron-3.5-Lightning-30B-A3B-BF16'"
          },
          {
            "event": "2026-08-11-nemotron-3-5-lightning",
            "date": "2026-08-11",
            "lab": "nvidia",
            "value": 10.8,
            "printed": "10.80",
            "source": "https://research.nvidia.com/labs/nemotron/files/NVIDIA-Nemotron-3.5-Super-and-Lightning-Technical-Report.pdf",
            "where": "Table 22 (PDF page 42), row 'HLE (no tools)', column 'BF16 acc.'"
          }
        ]
      },
      {
        "title": "HLE, with tools",
        "unit": "%",
        "rows": [
          {
            "event": "2025-04-16-o3-and-o4-mini",
            "date": "2025-04-16",
            "lab": "openai",
            "value": 24.9,
            "printed": "24.90",
            "source": "https://openai.com/index/introducing-o3-and-o4-mini",
            "where": "'What's changed' section, bar chart 'Humanity's Last Exam', 4th bar 'o3 (python + browsing** tools)' (page screenshot (job 20879 capture), part 02)"
          },
          {
            "event": "2025-04-16-o3-and-o4-mini",
            "date": "2025-04-16",
            "lab": "openai",
            "value": 24.3,
            "printed": "24.3%",
            "source": "https://openai.com/index/introducing-gpt-5",
            "where": "'Evaluations' section, bar chart 'Humanity's Last Exam (Full Set)*', 8th bar 'OpenAI o3 (python + browser)' (page screenshot (job 20879 capture), part 05)"
          },
          {
            "event": "2025-11-24-claude-opus-4-5",
            "date": "2025-11-24",
            "lab": "anthropic",
            "value": 43.2,
            "printed": "43.2%",
            "source": "https://www.anthropic.com/claude-opus-4-5-system-card",
            "where": "Figure 2.16.A 'Humanity's Last Exam (HLE)', panel 'With Search', bar 'Opus 4.5' (page 32)"
          },
          {
            "event": "2025-11-24-claude-opus-4-5",
            "date": "2025-11-24",
            "lab": "anthropic",
            "value": 43.4,
            "printed": "43.4%",
            "source": "https://www.anthropic.com/news/claude-opus-4-6",
            "where": "benchmark table (image), row 'Multidisciplinary reasoning, Humanity's Last Exam', column 'Opus 4.5', 'with tools'"
          },
          {
            "event": "2025-09-29-claude-sonnet-4-5",
            "date": "2025-09-29",
            "lab": "anthropic",
            "value": 28.4,
            "printed": "28.4%",
            "source": "https://www.anthropic.com/claude-opus-4-5-system-card",
            "where": "Figure 2.16.A 'Humanity's Last Exam (HLE)', panel 'With Search', bar 'Sonnet 4.5' (page 32)"
          },
          {
            "event": "2025-09-29-claude-sonnet-4-5",
            "date": "2025-09-29",
            "lab": "anthropic",
            "value": 33.6,
            "printed": "33.6%",
            "source": "https://www.anthropic.com/news/claude-opus-4-6",
            "where": "benchmark table (image), row 'Multidisciplinary reasoning, Humanity's Last Exam', column 'Sonnet 4.5', 'with tools'"
          },
          {
            "event": "2026-02-05-claude-opus-4-6",
            "date": "2026-02-05",
            "lab": "anthropic",
            "value": 53.0,
            "printed": "53.0%",
            "source": "https://www.anthropic.com/news/claude-opus-4-6",
            "where": "benchmark table (image), row 'Multidisciplinary reasoning, Humanity's Last Exam', column 'Opus 4.6', 'with tools'"
          },
          {
            "event": "2026-02-05-claude-opus-4-6",
            "date": "2026-02-05",
            "lab": "anthropic",
            "value": 53.3,
            "printed": "53.3%",
            "source": "https://www.anthropic.com/news/claude-opus-4-7",
            "where": "benchmark table (image), row 'Multidisciplinary reasoning, Humanity's Last Exam', column 'Opus 4.6', 'with tools'"
          },
          {
            "event": "2026-02-17-claude-sonnet-4-6",
            "date": "2026-02-17",
            "lab": "anthropic",
            "value": 49.0,
            "printed": "49.0%",
            "source": "https://www.anthropic.com/news/claude-sonnet-4-6",
            "where": "benchmark table (image), row 'Multidisciplinary reasoning, Humanity's Last Exam (HLE)', column 'Sonnet 4.6', 'with tools'"
          },
          {
            "event": "2026-02-17-claude-sonnet-4-6",
            "date": "2026-02-17",
            "lab": "anthropic",
            "value": 46.8,
            "printed": "46.8%",
            "source": "https://www.anthropic.com/news/claude-sonnet-5",
            "where": "benchmark table (image), row 'Multidisciplinary reasoning, Humanity's Last Exam', column 'Sonnet 4.6', 'with tools'"
          },
          {
            "event": "2026-04-02-qwen3-6-plus",
            "date": "2026-04-02",
            "lab": "alibaba",
            "value": 50.6,
            "printed": "50.6",
            "source": "https://qwen.ai/blog?id=qwen3.6",
            "where": "'Performance' section, 'Language' table, row 'HLE w/ tool', column 'Qwen3.6-Plus'"
          },
          {
            "event": "2026-04-02-qwen3-6-plus",
            "date": "2026-04-02",
            "lab": "alibaba",
            "value": 50.2,
            "printed": "50.2",
            "source": "https://qwen.ai/blog?id=qwen3.7",
            "where": "Qwen3.7 post, 'Performance' table, row 'HLE w/ tools', column 'Qwen3.6-Plus'"
          },
          {
            "event": "2026-06-30-claude-sonnet-5",
            "date": "2026-06-30",
            "lab": "anthropic",
            "value": 57.4,
            "printed": "57.4%",
            "source": "https://www.anthropic.com/news/claude-sonnet-5",
            "where": "benchmark table (image), row 'Multidisciplinary reasoning, Humanity's Last Exam', column 'Sonnet 5', 'with tools'"
          },
          {
            "event": "2026-06-30-claude-sonnet-5",
            "date": "2026-06-30",
            "lab": "anthropic",
            "value": 54.9,
            "printed": "54.9%",
            "source": "https://www.anthropic.com/claude-sonnet-5-5",
            "where": "benchmark table (HTML), row 'Multidisciplinary reasoning, Humanity's Last Exam', column 'Sonnet 5', 'with tools'"
          },
          {
            "event": "2026-06-09-claude-fable-5-mythos-5",
            "date": "2026-06-09",
            "lab": "anthropic",
            "value": 63.9,
            "printed": "63.9%",
            "source": "https://www.anthropic.com/news/claude-opus-5",
            "where": "benchmark table (image), row 'Multidisciplinary reasoning, Humanity's Last Exam', column 'Fable 5', 'with tools'"
          },
          {
            "event": "2026-06-09-claude-fable-5-mythos-5",
            "date": "2026-06-09",
            "lab": "anthropic",
            "value": 63.8,
            "printed": "63.8%",
            "source": "https://www.anthropic.com/claude-fable-and-mythos-5-1",
            "where": "benchmark table (HTML), row 'Multidisciplinary reasoning, Humanity's Last Exam', column 'Fable 5', 'with tools'"
          },
          {
            "event": "2026-07-24-claude-opus-5",
            "date": "2026-07-24",
            "lab": "anthropic",
            "value": 64.7,
            "printed": "64.7%",
            "source": "https://www.anthropic.com/news/claude-opus-5",
            "where": "benchmark table (image), row 'Multidisciplinary reasoning, Humanity's Last Exam', column 'Opus 5', 'with tools'"
          },
          {
            "event": "2026-07-24-claude-opus-5",
            "date": "2026-07-24",
            "lab": "anthropic",
            "value": 63.6,
            "printed": "63.6%",
            "source": "https://www.anthropic.com/claude-fable-and-mythos-5-1",
            "where": "benchmark table (HTML), row 'Multidisciplinary reasoning, Humanity's Last Exam', column 'Opus 5', 'with tools'"
          },
          {
            "event": "2026-09-01-claude-fable-5-1-mythos-5-1",
            "date": "2026-09-01",
            "lab": "anthropic",
            "value": 65.0,
            "printed": "65.0%",
            "source": "https://www.anthropic.com/claude-fable-and-mythos-5-1",
            "where": "benchmark table (HTML), row 'Multidisciplinary reasoning, Humanity's Last Exam', column 'Fable 5.1', 'with tools'"
          },
          {
            "event": "2026-09-01-claude-fable-5-1-mythos-5-1",
            "date": "2026-09-01",
            "lab": "anthropic",
            "value": 65.6,
            "printed": "65.6%",
            "source": "https://www.anthropic.com/claude-opus-5-5",
            "where": "benchmark table, row 'Multidisciplinary reasoning, Humanity's Last Exam', column 'Fable 5.1', 'with tools'"
          }
        ]
      },
      {
        "title": "LiveCodeBench",
        "unit": "%",
        "rows": [
          {
            "event": "2025-05-07-mistral-medium-3",
            "date": "2025-05-07",
            "lab": "mistral",
            "value": 30.3,
            "printed": "30.3%",
            "source": "https://mistral.ai/news/mistral-medium-3/",
            "where": "benchmark table image ('Table Medium 6'), section CODING, row 'LiveCodeBench (v6) 0-shot', column 'Mistral Medium 3'"
          },
          {
            "event": "2025-05-07-mistral-medium-3",
            "date": "2025-05-07",
            "lab": "mistral",
            "value": 29.1,
            "printed": "29.1%",
            "source": "https://mistral.ai/news/magistral/",
            "where": "later report: Magistral post, bar chart image ('Plot Magistral'), group 'LiveCodeBench (v5)', first bar (legend 'Mistral-Medium 3'), value label"
          },
          {
            "event": "2025-07-28-glm-4-5",
            "date": "2025-07-28",
            "lab": "zai",
            "value": 72.9,
            "printed": "72.9",
            "source": "https://z.ai/blog/glm-4.5",
            "where": "'Reasoning' section table, row 'LiveCodeBench 2407-2501', column 'GLM-4.5'"
          },
          {
            "event": "2025-07-28-glm-4-5",
            "date": "2025-07-28",
            "lab": "zai",
            "value": 63.3,
            "printed": "63.3",
            "source": "https://z.ai/blog/glm-4.6",
            "where": "GLM-4.6 post, benchmark chart image (eight panels), panel 'LiveCodeBench v6', second bar 'GLM-4.5', value label"
          },
          {
            "event": "2026-06-04-nemotron-3-ultra",
            "date": "2026-06-04",
            "lab": "nvidia",
            "value": 89.0,
            "printed": "89.0",
            "source": "https://research.nvidia.com/labs/nemotron/files/NVIDIA-Nemotron-3-Ultra-Technical-Report.pdf",
            "where": "Table 10 (PDF page 38), row 'LiveCodeBench (v6)', column 'N-3-Ultra'"
          },
          {
            "event": "2026-06-04-nemotron-3-ultra",
            "date": "2026-06-04",
            "lab": "nvidia",
            "value": 89.21,
            "printed": "89.21",
            "source": "https://research.nvidia.com/labs/nemotron/files/NVIDIA-Nemotron-3.5-Super-and-Lightning-Technical-Report.pdf",
            "where": "Table 10 (PDF page 23), row 'LiveCodeBench v6', column 'N3-Ultra 550B-A55B' (second column)"
          }
        ]
      },
      {
        "title": "MCP Atlas",
        "unit": "%",
        "rows": [
          {
            "event": "2026-02-12-glm-5",
            "date": "2026-02-12",
            "lab": "zai",
            "value": 67.8,
            "printed": "67.8",
            "source": "https://z.ai/blog/glm-5",
            "where": "benchmark table, row 'MCP-Atlas Public Set', column 'GLM-5 (Thinking)'"
          },
          {
            "event": "2026-02-12-glm-5",
            "date": "2026-02-12",
            "lab": "zai",
            "value": 69.2,
            "printed": "69.2",
            "source": "https://z.ai/blog/glm-5.1",
            "where": "benchmark table near the end, row 'MCP-Atlas Public Set', column 'GLM-5'"
          }
        ]
      },
      {
        "title": "MMLU-Pro",
        "unit": "%",
        "rows": [
          {
            "event": "2024-07-23-llama-3-1-405b",
            "date": "2024-07-23",
            "lab": "meta",
            "value": 73.3,
            "printed": "73.3",
            "source": "https://ai.meta.com/blog/meta-llama-3-1/",
            "where": "benchmark table (image), row 'MMLU PRO (5-shot, CoT)', column 'Llama 3.1 405B'"
          },
          {
            "event": "2024-07-23-llama-3-1-405b",
            "date": "2024-07-23",
            "lab": "meta",
            "value": 73.4,
            "printed": "73.4",
            "source": "https://ai.meta.com/blog/llama-4-multimodal-intelligence/",
            "where": "'Llama 4 Scout instruction-tuned benchmarks' table (image), row 'MMLU Pro', column 'Llama 3.1 405B'"
          }
        ]
      },
      {
        "title": "OSWorld 2.0, offline set, partial score",
        "unit": "%",
        "rows": [
          {
            "event": "2026-07-09-gpt-5-6",
            "date": "2026-07-09",
            "lab": "openai",
            "value": 62.6,
            "printed": "62.6%",
            "source": "https://openai.com/index/gpt-5-6",
            "where": "computer-use/agentic eval table (HTML), row 'OSWorld 2.0', column 'GPT‑5.6 Sol'"
          },
          {
            "event": "2026-07-09-gpt-5-6",
            "date": "2026-07-09",
            "lab": "openai",
            "value": 65.7,
            "printed": "65.7%",
            "source": "https://openai.com/index/gpt-6-astra",
            "where": "'Computer Use' table, row 'OSWorld 2.0 (v2026.08.08, offline set, partial score)', column 'GPT-5.6 Sol2'"
          },
          {
            "event": "2026-07-24-claude-opus-5",
            "date": "2026-07-24",
            "lab": "anthropic",
            "value": 70.6,
            "printed": "70.6%",
            "source": "https://www.anthropic.com/news/claude-opus-5",
            "where": "benchmark table image, row 'Computer use / OSWorld 2.0', column 'Opus 5'"
          },
          {
            "event": "2026-07-24-claude-opus-5",
            "date": "2026-07-24",
            "lab": "anthropic",
            "value": 75.4,
            "printed": "75.4%",
            "source": "https://www.anthropic.com/claude-fable-and-mythos-5-1",
            "where": "benchmark table (HTML), row 'Computer use / OSWorld 2.0 [2]', column 'Opus 5', the value tagged 'partial'"
          }
        ]
      },
      {
        "title": "SWE-bench Pro",
        "unit": "%",
        "rows": [
          {
            "event": "2026-04-08-muse-spark",
            "date": "2026-04-08",
            "lab": "meta",
            "value": 52.4,
            "printed": "52.4",
            "source": "https://ai.meta.com/blog/introducing-muse-spark-msl/",
            "where": "main benchmark table (image), row 'SWE-Bench Pro', column 'Muse Spark Thinking'"
          },
          {
            "event": "2026-04-08-muse-spark",
            "date": "2026-04-08",
            "lab": "meta",
            "value": 55.0,
            "printed": "55.0",
            "source": "https://ai.meta.com/blog/introducing-muse-spark-meta-model-api/",
            "where": "Muse Spark 1.1 benchmark table (image), row 'SWE-Bench Pro', column 'Muse Spark'"
          },
          {
            "event": "2026-04-23-gpt-5-5",
            "date": "2026-04-23",
            "lab": "openai",
            "value": 58.6,
            "printed": "58.6%",
            "source": "https://openai.com/index/introducing-gpt-5-5",
            "where": "coding eval table, row 'SWE-Bench Pro (Public) *', column 'GPT‑5.5'"
          },
          {
            "event": "2026-04-23-gpt-5-5",
            "date": "2026-04-23",
            "lab": "openai",
            "value": 59.4,
            "printed": "59.4%",
            "source": "https://openai.com/index/gpt-5-6",
            "where": "coding table (columns GPT-5.6 Sol, Sol Ultra, Terra, Luna, GPT-5.5, ...), row 'SWE-Bench Pro', column 'GPT-5.5'"
          },
          {
            "event": "2026-06-01-qwen3-7-plus",
            "date": "2026-06-01",
            "lab": "alibaba",
            "value": 57.6,
            "printed": "57.6",
            "source": "https://qwen.ai/blog?id=qwen3.7-plus",
            "where": "language benchmark table (first table), Coding Agent block, row 'SWE-Pro', column 'Qwen3.7-Plus'"
          },
          {
            "event": "2026-06-01-qwen3-7-plus",
            "date": "2026-06-01",
            "lab": "alibaba",
            "value": 55.8,
            "printed": "55.8",
            "source": "https://qwen.ai/blog?id=qwen3.8-flash-next",
            "where": "later report: Qwen3.8-Flash-Next post, Language table (first table), row 'Agentic coding / SWE-bench Pro', column 'Qwen3.7-Plus'"
          }
        ]
      },
      {
        "title": "SWE-bench Verified",
        "unit": "%",
        "rows": [
          {
            "event": "2024-09-12-openai-o1",
            "date": "2024-09-12",
            "lab": "openai",
            "value": 41.4,
            "printed": "41.4%",
            "source": "https://cdn.openai.com/o1-preview-system-card-20240917.pdf",
            "where": "section 4.5.2 'SWE-bench Verified', paragraph under the 'SWE-Bench Verified (with 5-try Agentless Scaffold)' chart, sentence on the final o1-preview model"
          },
          {
            "event": "2024-09-12-openai-o1",
            "date": "2024-09-12",
            "lab": "openai",
            "value": 41.3,
            "printed": "41.3",
            "source": "https://openai.com/index/o1-and-new-tools-for-developers",
            "where": "eval table (Category / Eval / o1-2024-12-17 / o1-preview), row 'Coding: SWE-bench Verified', column 'o1-preview'"
          },
          {
            "event": "2025-08-07-gpt-5",
            "date": "2025-08-07",
            "lab": "openai",
            "value": 74.9,
            "printed": "74.9%",
            "source": "https://openai.com/index/introducing-gpt-5",
            "where": "text, opening summary of results"
          },
          {
            "event": "2025-08-07-gpt-5",
            "date": "2025-08-07",
            "lab": "openai",
            "value": 72.8,
            "printed": "72.8%",
            "source": "https://openai.com/index/gpt-5-1-for-developers",
            "where": "evaluation table, row 'SWE-bench Verified (all 500 problems)', column 'GPT-5 (high)'"
          },
          {
            "event": "2025-10-27-minimax-m2",
            "date": "2025-10-27",
            "lab": "minimax",
            "value": 69.4,
            "printed": "69.4",
            "source": "https://www.minimax.io/blog/minimax-m2-en-1748600000",
            "where": "first benchmark chart image (wide, eight panels), panel 'SWE-bench Verified', first bar 'MiniMax-M2' (red), value label"
          },
          {
            "event": "2025-10-27-minimax-m2",
            "date": "2025-10-27",
            "lab": "minimax",
            "value": 68.1,
            "printed": "68.1",
            "source": "https://www.minimax.io/news/minimax-m21",
            "where": "'Benchmarks' section, second (agent frameworks and specialized benchmarks) table image, row 'SWE-bench Verified (Droid)', column 'MiniMax-M2'"
          },
          {
            "event": "2026-02-16-qwen3-5",
            "date": "2026-02-16",
            "lab": "alibaba",
            "value": 76.4,
            "printed": "76.4",
            "source": "https://qwen.ai/blog?id=qwen3.5",
            "where": "'Performance' section, 'Language' table, row 'SWE-bench Verified', column 'Qwen3.5-397B-A17B'"
          },
          {
            "event": "2026-02-16-qwen3-5",
            "date": "2026-02-16",
            "lab": "alibaba",
            "value": 76.2,
            "printed": "76.2",
            "source": "https://qwen.ai/blog?id=qwen3.6",
            "where": "Qwen3.6-Plus post, 'Performance' section, 'Language' table, row 'SWE-bench Verified', column 'Qwen3.5-397B-A17B'"
          },
          {
            "event": "2026-06-04-nemotron-3-ultra",
            "date": "2026-06-04",
            "lab": "nvidia",
            "value": 71.9,
            "printed": "71.9",
            "source": "https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4/blob/02c7d9e60f328a4fb9d990d6b613057b19ab8a8f/README.md",
            "where": "benchmark table, row 'SWE-Bench Verified', column 'Nemotron 3 Ultra BF16'"
          },
          {
            "event": "2026-06-04-nemotron-3-ultra",
            "date": "2026-06-04",
            "lab": "nvidia",
            "value": 73.93,
            "printed": "73.93",
            "source": "https://research.nvidia.com/labs/nemotron/files/NVIDIA-Nemotron-3.5-Super-and-Lightning-Technical-Report.pdf",
            "where": "Table 10 (PDF page 23), row 'SWE-bench Verified', column 'N3-Ultra 550B-A55B' (second column)"
          }
        ]
      },
      {
        "title": "Terminal-Bench 2.0",
        "unit": "%",
        "rows": [
          {
            "event": "2025-11-18-gemini-3-pro",
            "date": "2025-11-18",
            "lab": "google",
            "value": 54.2,
            "printed": "54.2%",
            "source": "https://blog.google/innovation-and-ai/technology/developers-tools/gemini-3-developers/",
            "where": "benchmark table (image), row 'Terminal-Bench 2.0', 'Terminus-2 agent', column 'Gemini 3 Pro'"
          },
          {
            "event": "2025-11-18-gemini-3-pro",
            "date": "2025-11-18",
            "lab": "google",
            "value": 56.9,
            "printed": "56.9%",
            "source": "https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-1-pro/",
            "where": "benchmark table (image), row 'Terminal-Bench 2.0', 'Terminus-2 harness', column 'Gemini 3 Pro, Thinking (High)'"
          },
          {
            "event": "2025-11-24-claude-opus-4-5",
            "date": "2025-11-24",
            "lab": "anthropic",
            "value": 59.3,
            "printed": "59.3%",
            "source": "https://www.anthropic.com/news/claude-opus-4-5",
            "where": "benchmark table (image), row 'Agentic terminal coding, Terminal-bench 2.0', column 'Opus 4.5'"
          },
          {
            "event": "2025-11-24-claude-opus-4-5",
            "date": "2025-11-24",
            "lab": "anthropic",
            "value": 59.8,
            "printed": "59.8%",
            "source": "https://www.anthropic.com/claude-opus-4-6-system-card",
            "where": "Table 2.3.A (PDF page 18), row 'Terminal-Bench 2.0', column 'Claude Opus 4.5'"
          },
          {
            "event": "2025-09-29-claude-sonnet-4-5",
            "date": "2025-09-29",
            "lab": "anthropic",
            "value": 50.0,
            "printed": "50.0%",
            "source": "https://www.anthropic.com/news/claude-opus-4-5",
            "where": "benchmark table (image), row 'Agentic terminal coding, Terminal-bench 2.0', column 'Sonnet 4.5'"
          },
          {
            "event": "2025-09-29-claude-sonnet-4-5",
            "date": "2025-09-29",
            "lab": "anthropic",
            "value": 51.0,
            "printed": "51.0%",
            "source": "https://www.anthropic.com/claude-opus-4-6-system-card",
            "where": "Table 2.3.A (PDF page 18), row 'Terminal-Bench 2.0', column 'Claude Sonnet 4.5'"
          },
          {
            "event": "2025-12-22-glm-4-7",
            "date": "2025-12-22",
            "lab": "zai",
            "value": 41.0,
            "printed": "41.0",
            "source": "https://z.ai/blog/glm-4.7",
            "where": "benchmark table, row 'Terminal Bench 2.0', column 'GLM-4.7'"
          },
          {
            "event": "2025-12-22-glm-4-7",
            "date": "2025-12-22",
            "lab": "zai",
            "value": 32.8,
            "printed": "32.8",
            "source": "https://z.ai/blog/glm-5",
            "where": "benchmark table, row 'Terminal-Bench 2.0 Claude Code', column 'GLM-4.7 (Thinking)'"
          }
        ]
      },
      {
        "title": "Terminal-Bench 2.1",
        "unit": "%",
        "rows": [
          {
            "event": "2026-05-28-claude-opus-4-8",
            "date": "2026-05-28",
            "lab": "anthropic",
            "value": 74.6,
            "printed": "74.6%",
            "source": "https://www.anthropic.com/news/claude-opus-4-8",
            "where": "benchmark table image, row 'Agentic terminal coding / Terminal-Bench 2.1', column 'Opus 4.8'"
          },
          {
            "event": "2026-05-28-claude-opus-4-8",
            "date": "2026-05-28",
            "lab": "anthropic",
            "value": 82.7,
            "printed": "82.7%",
            "source": "https://www.anthropic.com/news/claude-sonnet-5",
            "where": "benchmark table (image), row 'Agentic coding, Terminal-Bench 2.1', column 'Opus 4.8 (For reference)'"
          },
          {
            "event": "2026-06-04-nemotron-3-ultra",
            "date": "2026-06-04",
            "lab": "nvidia",
            "value": 56.4,
            "printed": "56.4",
            "source": "https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4/blob/02c7d9e60f328a4fb9d990d6b613057b19ab8a8f/README.md",
            "where": "benchmark table, row 'Terminal Bench 2.1', column 'Nemotron 3 Ultra BF16'"
          },
          {
            "event": "2026-06-04-nemotron-3-ultra",
            "date": "2026-06-04",
            "lab": "nvidia",
            "value": 54.78,
            "printed": "54.78",
            "source": "https://research.nvidia.com/labs/nemotron/files/NVIDIA-Nemotron-3.5-Super-and-Lightning-Technical-Report.pdf",
            "where": "Table 10 (PDF page 23), row 'Terminal-Bench 2.1', column 'N3-Ultra 550B-A55B' (second column)"
          },
          {
            "event": "2026-07-09-muse-spark-1-1",
            "date": "2026-07-09",
            "lab": "meta",
            "value": 80.0,
            "printed": "80.0",
            "source": "https://ai.meta.com/blog/introducing-muse-spark-meta-model-api/",
            "where": "benchmark table (image), row 'Terminal-Bench 2.1', column 'Muse Spark 1.1'"
          },
          {
            "event": "2026-07-09-muse-spark-1-1",
            "date": "2026-07-09",
            "lab": "meta",
            "value": 76.2,
            "printed": "76.2%",
            "source": "https://research.meta.ai/blog/introducing-muse-code-and-muse-spark-1-2",
            "where": "bar chart 'Terminal-Bench 2.1' (image, Coding tab), value label above the bar 'Muse Spark 1.1, mini-swe-agent'"
          }
        ]
      },
      {
        "title": "TB-Science 0.1",
        "unit": "%",
        "rows": [
          {
            "event": "2026-09-03-gpt-6-astra",
            "date": "2026-09-03",
            "lab": "openai",
            "value": 64.6,
            "printed": "64.6%",
            "source": "https://openai.com/index/gpt-6-astra",
            "where": "Academic table, row 'Terminal-Bench Science 0.1', column 'GPT‑6 Astra'"
          },
          {
            "event": "2026-09-03-gpt-6-astra",
            "date": "2026-09-03",
            "lab": "openai",
            "value": 68.1,
            "printed": "68.1%",
            "source": "https://openai.com/index/introducing-gpt-6-1-sol",
            "where": "'Scientific research' section, paragraph after the Terminal-Bench Science 0.1 cost sentence: 'GPT-6 Astra still achieves the highest score among the models tested at ...'"
          }
        ]
      }
    ],
    "sources": [
      "https://ai.meta.com/blog/introducing-muse-spark-meta-model-api/",
      "https://ai.meta.com/blog/introducing-muse-spark-msl/",
      "https://ai.meta.com/blog/llama-4-multimodal-intelligence/",
      "https://ai.meta.com/blog/meta-llama-3-1/",
      "https://api-docs.deepseek.com/news/news260813/",
      "https://api-docs.deepseek.com/news/news260910/",
      "https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-1-pro/",
      "https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-6-flash-3-5-flash-lite-3-5-flash-cyber/",
      "https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-3-7-flash/",
      "https://blog.google/innovation-and-ai/models-and-research/google-deepmind/gemini-model-thinking-updates-march-2025/",
      "https://blog.google/innovation-and-ai/technology/developers-tools/gemini-3-developers/",
      "https://cdn.openai.com/o1-preview-system-card-20240917.pdf",
      "https://developer.nvidia.com/blog/build-more-accurate-and-efficient-ai-agents-with-the-new-nvidia-llama-nemotron-super-v1-5/",
      "https://huggingface.co/nvidia/Llama-3_3-Nemotron-Super-49B-v1/blob/9314922284e3ef0e86ddbcad0d8b8dbd62238cfb/README.md",
      "https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-FP8/blob/fa5699ff0bb9da723217af52488c0f9653512ffb/README.md",
      "https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-FP8/blob/1408abd63ba6ba1c37a78b94d332dac35da45370/README.md",
      "https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4/blob/02c7d9e60f328a4fb9d990d6b613057b19ab8a8f/README.md",
      "https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16/blob/942b56629bd9b52c6813bb5f68bb5d10ba8cca68/README.md",
      "https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-NVFP4/blob/b14872a58afd94e2fbbc1e5a1fe4ebb7dfdc5fdd/README.md",
      "https://mistral.ai/news/magistral/",
      "https://mistral.ai/news/mistral-medium-3/",
      "https://openai.com/index/gpt-4-1",
      "https://openai.com/index/gpt-5-1-for-developers",
      "https://openai.com/index/gpt-5-6",
      "https://openai.com/index/gpt-6-astra",
      "https://openai.com/index/hello-gpt-4o",
      "https://openai.com/index/introducing-deep-research",
      "https://openai.com/index/introducing-gpt-4-5",
      "https://openai.com/index/introducing-gpt-5",
      "https://openai.com/index/introducing-gpt-5-4",
      "https://openai.com/index/introducing-gpt-5-4-mini-and-nano",
      "https://openai.com/index/introducing-gpt-5-5",
      "https://openai.com/index/introducing-gpt-5-for-developers",
      "https://openai.com/index/introducing-gpt-6-1-sol",
      "https://openai.com/index/introducing-gpt-6-sol-and-luna",
      "https://openai.com/index/introducing-o3-and-o4-mini",
      "https://openai.com/index/learning-to-reason-with-llms",
      "https://openai.com/index/o1-and-new-tools-for-developers",
      "https://openai.com/index/openai-o1-mini-advancing-cost-efficient-reasoning",
      "https://openai.com/index/openai-o3-mini",
      "https://qwen.ai/blog?id=qwen3.5",
      "https://qwen.ai/blog?id=qwen3.6",
      "https://qwen.ai/blog?id=qwen3.7",
      "https://qwen.ai/blog?id=qwen3.7-plus",
      "https://qwen.ai/blog?id=qwen3.8-flash-next",
      "https://research.meta.ai/blog/introducing-muse-code-and-muse-spark-1-2",
      "https://research.meta.ai/blog/introducing-muse-spark-1-3",
      "https://research.nvidia.com/labs/nemotron/files/NVFP4-QAD-Report.pdf",
      "https://research.nvidia.com/labs/nemotron/files/NVIDIA-Nemotron-3-Nano-Technical-Report.pdf",
      "https://research.nvidia.com/labs/nemotron/files/NVIDIA-Nemotron-3-Super-Technical-Report.pdf",
      "https://research.nvidia.com/labs/nemotron/files/NVIDIA-Nemotron-3-Ultra-Technical-Report.pdf",
      "https://research.nvidia.com/labs/nemotron/files/NVIDIA-Nemotron-3.5-Bolt-74B-A9B-24GB-Technical-Report.pdf",
      "https://research.nvidia.com/labs/nemotron/files/NVIDIA-Nemotron-3.5-Super-and-Lightning-Technical-Report.pdf",
      "https://www.anthropic.com/claude-fable-5-1-mythos-5-1-system-card",
      "https://www.anthropic.com/claude-fable-5-mythos-5-system-card",
      "https://www.anthropic.com/claude-fable-and-mythos-5-1",
      "https://www.anthropic.com/claude-opus-4-5-system-card",
      "https://www.anthropic.com/claude-opus-4-6-system-card",
      "https://www.anthropic.com/claude-opus-4-8-system-card",
      "https://www.anthropic.com/claude-opus-5-5",
      "https://www.anthropic.com/claude-opus-5-5-system-card",
      "https://www.anthropic.com/claude-sonnet-5-5",
      "https://www.anthropic.com/claude-sonnet-5-5-system-card",
      "https://www.anthropic.com/claude-sonnet-5-system-card",
      "https://www.anthropic.com/news/claude-opus-4-1",
      "https://www.anthropic.com/news/claude-opus-4-5",
      "https://www.anthropic.com/news/claude-opus-4-6",
      "https://www.anthropic.com/news/claude-opus-4-7",
      "https://www.anthropic.com/news/claude-opus-4-8",
      "https://www.anthropic.com/news/claude-opus-5",
      "https://www.anthropic.com/news/claude-sonnet-4-6",
      "https://www.anthropic.com/news/claude-sonnet-5",
      "https://www.minimax.io/blog/minimax-m2-en-1748600000",
      "https://www.minimax.io/news/minimax-m21",
      "https://z.ai/blog/glm-4.5",
      "https://z.ai/blog/glm-4.6",
      "https://z.ai/blog/glm-4.7",
      "https://z.ai/blog/glm-5",
      "https://z.ai/blog/glm-5.1"
    ]
  }
}