{
  "api_version": "1",
  "updated": "2026-10-26",
  "latest_event": "2026-10-24",
  "rights": "© Frontier Log, all rights reserved.",
  "model": {
    "slug": "kimi-k3",
    "name": "Kimi K3",
    "lab": "moonshot",
    "date": "2026-07-16",
    "flagship": false,
    "event": "https://frontierlog.ai/event/2026-07-16-kimi-k3",
    "releases": [],
    "variant_of": null,
    "made_by": "Moonshot AI",
    "url": "https://frontierlog.ai/model/kimi-k3",
    "results": [
      {
        "id": "kimi-k3:hle:no-tools:1",
        "model": "kimi-k3",
        "model_name": "Kimi K3",
        "model_date": "2026-07-16",
        "event": "https://frontierlog.ai/event/2026-07-16-kimi-k3",
        "benchmark": "hle",
        "benchmark_name": "Humanity's Last Exam",
        "setting": "no-tools",
        "setting_name": "No tools",
        "score": 43.5,
        "unit": "%",
        "printed": "43.5",
        "reported_by": "moonshot",
        "run_by": null,
        "label": "Kimi K3 (max)",
        "source": "https://www.kimi.ai/blog/kimi-k3",
        "source_title": "Kimi K3: Open Frontier Intelligence",
        "source_date": "2026-07-16",
        "where": "'Full Benchmark Table', row 'HLE-Full', column 'Kimi K3 (max)'",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "full set; max reasoning effort",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-17",
        "check": "human",
        "corroborated": true,
        "corroborated_on": "2026-10-19",
        "corroborated_by": [
          {
            "url": "https://api-docs.deepseek.com/news/news260910/",
            "name": "DeepSeek-V4.1-Flash",
            "title": "DeepSeek-V4.1-Flash: Smarter, Faster, More Efficient",
            "lab": "deepseek",
            "date": "2026-09-10",
            "printed": "43.5",
            "where": "Benchmark table image; row 'HLE'; column 'Kimi K3'; separate 'HLE (w/tools)' row exists"
          },
          {
            "url": "https://api-docs.deepseek.com/news/news261017/",
            "name": "DeepSeek-V4.1-Flash-1017",
            "title": "DeepSeek-V4.1-Flash-1017: Wish-Driven Programming Made Real",
            "lab": "deepseek",
            "date": "2026-10-17",
            "printed": "43.5",
            "where": "Benchmark table image; row 'HLE (w/o tools)'; column 'Kimi K3'; first value of 'x/y*' cells is the full set"
          }
        ],
        "source_capture": "https://frontierlog.ai/captures/fd6adf73f1b281b7.png"
      },
      {
        "id": "kimi-k3:browsecomp:default:1",
        "model": "kimi-k3",
        "model_name": "Kimi K3",
        "model_date": "2026-07-16",
        "event": "https://frontierlog.ai/event/2026-07-16-kimi-k3",
        "benchmark": "browsecomp",
        "benchmark_name": "BrowseComp",
        "setting": "default",
        "setting_name": "",
        "score": 91.2,
        "unit": "%",
        "printed": "91.2",
        "reported_by": "moonshot",
        "run_by": null,
        "label": "Kimi K3 (max)",
        "source": "https://www.kimi.ai/blog/kimi-k3",
        "source_title": "Kimi K3: Open Frontier Intelligence",
        "source_date": "2026-07-16",
        "where": "'Full Benchmark Table', row 'BrowseComp', column 'Kimi K3 (max)'",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "max reasoning effort; context compaction at 300K tokens",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-17",
        "check": "human",
        "corroborated": false,
        "corroborated_on": null,
        "corroborated_by": [],
        "source_capture": "https://frontierlog.ai/captures/fd6adf73f1b281b7.png"
      },
      {
        "id": "kimi-k3:browsecomp:default:2",
        "model": "kimi-k3",
        "model_name": "Kimi K3",
        "model_date": "2026-07-16",
        "event": "https://frontierlog.ai/event/2026-07-16-kimi-k3",
        "benchmark": "browsecomp",
        "benchmark_name": "BrowseComp",
        "setting": "default",
        "setting_name": "",
        "score": 90.4,
        "unit": "%",
        "printed": "90.4",
        "reported_by": "moonshot",
        "run_by": null,
        "label": "Kimi K3",
        "source": "https://www.kimi.ai/blog/kimi-k3",
        "source_title": "Kimi K3: Open Frontier Intelligence",
        "source_date": "2026-07-16",
        "where": "Footnotes, 'Productivity and agentic benchmarks', the 'BrowseComp.' note (score with a 1M-token context window and no context management)",
        "own": true,
        "primary": false,
        "parallel": false,
        "notes": "1M-token context, no context management",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-17",
        "check": "human",
        "corroborated": false,
        "corroborated_on": null,
        "corroborated_by": []
      },
      {
        "id": "kimi-k3:deepswe-1-1:default:1",
        "model": "kimi-k3",
        "model_name": "Kimi K3",
        "model_date": "2026-07-16",
        "event": "https://frontierlog.ai/event/2026-07-16-kimi-k3",
        "benchmark": "deepswe-1-1",
        "benchmark_name": "DeepSWE v1.1",
        "setting": "default",
        "setting_name": "",
        "score": 67.5,
        "unit": "%",
        "printed": "67.5",
        "reported_by": "moonshot",
        "run_by": null,
        "label": "Kimi K3 (max)",
        "source": "https://www.kimi.ai/blog/kimi-k3",
        "source_title": "Kimi K3: Open Frontier Intelligence",
        "source_date": "2026-07-16",
        "where": "'Full Benchmark Table', row 'DeepSWE' (Coding), column 'Kimi K3 (max)'",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "Kimi Code harness, max effort",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-17",
        "check": "human",
        "corroborated": true,
        "corroborated_on": "2026-10-19",
        "corroborated_by": [
          {
            "url": "https://api-docs.deepseek.com/news/news260910/",
            "name": "DeepSeek-V4.1-Flash",
            "title": "DeepSeek-V4.1-Flash: Smarter, Faster, More Efficient",
            "lab": "deepseek",
            "date": "2026-09-10",
            "printed": "67.5",
            "where": "Bar chart image; group 'DeepSWE v1.1'; value label on Kimi-K3 bar"
          },
          {
            "url": "https://api-docs.deepseek.com/news/news261017/",
            "name": "DeepSeek-V4.1-Flash-1017",
            "title": "DeepSeek-V4.1-Flash-1017: Wish-Driven Programming Made Real",
            "lab": "deepseek",
            "date": "2026-10-17",
            "printed": "67.5",
            "where": "Benchmark table image; row 'DeepSWE v1.1'; column 'Kimi K3'; footnote: Mini-SWE-V2 framework"
          },
          {
            "url": "https://z.ai/blog/glm-5.3",
            "name": "GLM-5.3",
            "title": "GLM-5.3: Frontier Coding with Emergent Cyber Capabilities",
            "lab": "zai",
            "date": "2026-08-14",
            "printed": "67.5",
            "where": "'Performance across comparison models' table, row 'DeepSWE v1.1', column 'Kimi K3'"
          }
        ],
        "source_capture": "https://frontierlog.ai/captures/fd6adf73f1b281b7.png"
      },
      {
        "id": "kimi-k3:deepswe-1-1:default:2",
        "model": "kimi-k3",
        "model_name": "Kimi K3",
        "model_date": "2026-07-16",
        "event": "https://frontierlog.ai/event/2026-07-16-kimi-k3",
        "benchmark": "deepswe-1-1",
        "benchmark_name": "DeepSWE v1.1",
        "setting": "default",
        "setting_name": "",
        "score": 67.3,
        "unit": "%",
        "printed": "67.3",
        "reported_by": "moonshot",
        "run_by": null,
        "label": "Kimi K3",
        "source": "https://www.kimi.ai/blog/kimi-k3",
        "source_title": "Kimi K3: Open Frontier Intelligence",
        "source_date": "2026-07-16",
        "where": "Footnotes, 'Coding benchmarks', the 'DeepSWE.' note (score on the official DeepSWE leaderboard with the mini-SWE-agent harness)",
        "own": true,
        "primary": false,
        "parallel": false,
        "notes": "mini-SWE-agent harness, as on the official leaderboard",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-17",
        "check": "human",
        "corroborated": false,
        "corroborated_on": null,
        "corroborated_by": []
      },
      {
        "id": "kimi-k3:mcp-atlas:default:1",
        "model": "kimi-k3",
        "model_name": "Kimi K3",
        "model_date": "2026-07-16",
        "event": "https://frontierlog.ai/event/2026-07-16-kimi-k3",
        "benchmark": "mcp-atlas",
        "benchmark_name": "MCP Atlas",
        "setting": "default",
        "setting_name": "",
        "score": 84.2,
        "unit": "%",
        "printed": "84.2",
        "reported_by": "moonshot",
        "run_by": null,
        "label": "Kimi K3 (max)",
        "source": "https://www.kimi.ai/blog/kimi-k3",
        "source_title": "Kimi K3: Open Frontier Intelligence",
        "source_date": "2026-07-16",
        "where": "'Full Benchmark Table', row 'MCP Atlas', column 'Kimi K3 (max)'",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "max reasoning effort; 500-task public set; 100-turn limit; Gemini 3.1 Pro judge",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-18",
        "check": "human",
        "corroborated": false,
        "corroborated_on": null,
        "corroborated_by": [],
        "source_capture": "https://frontierlog.ai/captures/fd6adf73f1b281b7.png"
      },
      {
        "id": "kimi-k3:automationbench:default:1",
        "model": "kimi-k3",
        "model_name": "Kimi K3",
        "model_date": "2026-07-16",
        "event": "https://frontierlog.ai/event/2026-07-16-kimi-k3",
        "benchmark": "automationbench",
        "benchmark_name": "AutomationBench",
        "setting": "default",
        "setting_name": "",
        "score": 30.8,
        "unit": "%",
        "printed": "30.8",
        "reported_by": "moonshot",
        "run_by": null,
        "label": "Kimi K3 (max)",
        "source": "https://www.kimi.ai/blog/kimi-k3",
        "source_title": "Kimi K3: Open Frontier Intelligence",
        "source_date": "2026-07-16",
        "where": "'Full Benchmark Table', row 'Automation Bench' (Agentic), column 'Kimi K3 (max)'",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "600-task public subset; max reasoning effort",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-20",
        "check": "human",
        "corroborated": false,
        "corroborated_on": null,
        "corroborated_by": [],
        "source_capture": "https://frontierlog.ai/captures/fd6adf73f1b281b7.png"
      },
      {
        "id": "kimi-k3:gpqa-diamond:default:1",
        "model": "kimi-k3",
        "model_name": "Kimi K3",
        "model_date": "2026-07-16",
        "event": "https://frontierlog.ai/event/2026-07-16-kimi-k3",
        "benchmark": "gpqa-diamond",
        "benchmark_name": "GPQA Diamond",
        "setting": "default",
        "setting_name": "",
        "score": 93.5,
        "unit": "%",
        "printed": "93.5",
        "reported_by": "moonshot",
        "run_by": null,
        "label": "Kimi K3 (max)",
        "source": "https://www.kimi.ai/blog/kimi-k3",
        "source_title": "Kimi K3: Open Frontier Intelligence",
        "source_date": "2026-07-16",
        "where": "'Full Benchmark Table', row 'GPQA-Diamond' (Reasoning & Knowledge), column 'Kimi K3 (max)'",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "max reasoning effort",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-20",
        "check": "human",
        "corroborated": false,
        "corroborated_on": null,
        "corroborated_by": [],
        "source_capture": "https://frontierlog.ai/captures/fd6adf73f1b281b7.png"
      },
      {
        "id": "kimi-k3:hle:with-tools:1",
        "model": "kimi-k3",
        "model_name": "Kimi K3",
        "model_date": "2026-07-16",
        "event": "https://frontierlog.ai/event/2026-07-16-kimi-k3",
        "benchmark": "hle",
        "benchmark_name": "Humanity's Last Exam",
        "setting": "with-tools",
        "setting_name": "With tools",
        "score": 56.0,
        "unit": "%",
        "printed": "56.0",
        "reported_by": "moonshot",
        "run_by": null,
        "label": "Kimi K3 (max)",
        "source": "https://www.kimi.ai/blog/kimi-k3",
        "source_title": "Kimi K3: Open Frontier Intelligence",
        "source_date": "2026-07-16",
        "where": "'Full Benchmark Table', row 'HLE-Full w/ tools' (Reasoning & Knowledge), column 'Kimi K3 (max)'",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "full set; max reasoning effort",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-20",
        "check": "human",
        "corroborated": false,
        "corroborated_on": null,
        "corroborated_by": [],
        "source_capture": "https://frontierlog.ai/captures/fd6adf73f1b281b7.png"
      }
    ]
  }
}