{
  "api_version": "1",
  "updated": "2026-10-30",
  "latest_event": "2026-10-29",
  "rights": "© Frontier Log, all rights reserved.",
  "benchmark": {
    "slug": "sec-bench-pro-260505",
    "name": "SEC-bench Pro 260505",
    "family": "sec-bench-pro",
    "unit": "%",
    "higher_is_better": true,
    "status": "superseded",
    "active": false,
    "superseded_by": null,
    "retired": null,
    "benchmark_url": "https://sec-bench.github.io/260505/",
    "introduced": "2026-05-05",
    "creator": "University of Illinois Urbana-Champaign",
    "introduced_source": "https://sec-bench.github.io/260505/",
    "settings": [
      {
        "setting": "default",
        "name": "",
        "url": "https://frontierlog.ai/benchmark/sec-bench-pro-260505",
        "results": 4
      }
    ]
  },
  "total": 4,
  "results": [
    {
      "id": "deepseek-v4-pro-0813:sec-bench-pro-260505:default:1",
      "model": "deepseek-v4-pro-0813",
      "model_name": "DeepSeek-V4-Pro-0813",
      "model_date": "2026-08-13",
      "event": "https://frontierlog.ai/event/2026-08-13-deepseek-v4-pro-ga",
      "benchmark": "sec-bench-pro-260505",
      "benchmark_name": "SEC-bench Pro 260505",
      "setting": "default",
      "setting_name": "",
      "score": 56.4,
      "unit": "%",
      "printed": "56.4",
      "reported_by": "deepseek",
      "run_by": null,
      "label": "DeepSeek V4-Pro 0813",
      "source": "https://api-docs.deepseek.com/news/news260910/",
      "source_title": "DeepSeek-V4.1-Flash: Smarter, Faster, More Efficient",
      "source_date": "2026-09-10",
      "where": "benchmark table image (v4.1_260910_benchmark_table_en), row 'SEC-Bench Pro', column 'DeepSeek V4-Pro 0813'",
      "own": false,
      "primary": true,
      "parallel": false,
      "notes": "max reasoning effort",
      "caveat": "",
      "human_verified": false,
      "human_verified_on": null,
      "check": "corroborated",
      "corroborated": true,
      "corroborated_on": "2026-10-30",
      "corroborated_by": [
        {
          "url": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash",
          "name": "DeepSeek-V4.1-Flash card",
          "title": "deepseek-ai/DeepSeek-V4.1-Flash model card",
          "lab": "deepseek",
          "date": "2026-09-10",
          "printed": "56.4",
          "where": "README, 'Comparison with frontier models (Max reasoning effort)' table, row 'SEC-Bench Pro (Pass@1)', column 'DS-V4-Pro' (no 0813 suffix)"
        },
        {
          "url": "https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-4-carbon/",
          "name": "Gemini 4 Carbon",
          "title": "Gemini 4 Carbon: our next frontier model",
          "lab": "google",
          "date": "2026-10-28",
          "printed": "56.4%",
          "where": "SEC-bench Pro bar chart (Pass@1), value label on bar 'DeepSeek-V4 -Pro' (no 0813 suffix)"
        },
        {
          "url": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash/blob/main/DeepSeek_V41_Tech_Report.pdf",
          "name": "DeepSeek-V4.1-Flash card",
          "title": "DeepSeek-V4.1-Flash: Pushing the Limits of KV Cache Compression",
          "lab": "deepseek",
          "date": "2026-09-10",
          "printed": "56.4",
          "where": "Table 3, row 'SEC-Bench Pro (Pass@1)', column 'DS-V4-Pro' (Max; no 0813 suffix)"
        }
      ],
      "source_image": "https://frontierlog.ai/sources/19d3b83b1b9e5f5b.png",
      "source_image_original": "https://api-docs.deepseek.com/img/v4.1_260910_benchmark_table_en.png"
    },
    {
      "id": "deepseek-v4-1-flash:sec-bench-pro-260505:default:1",
      "model": "deepseek-v4-1-flash",
      "model_name": "DeepSeek-V4.1-Flash",
      "model_date": "2026-09-10",
      "event": "https://frontierlog.ai/event/2026-09-10-deepseek-v4-1-flash",
      "benchmark": "sec-bench-pro-260505",
      "benchmark_name": "SEC-bench Pro 260505",
      "setting": "default",
      "setting_name": "",
      "score": 62.8,
      "unit": "%",
      "printed": "62.8",
      "reported_by": "deepseek",
      "run_by": null,
      "label": "DeepSeek V4.1-Flash",
      "source": "https://api-docs.deepseek.com/news/news260910/",
      "source_title": "DeepSeek-V4.1-Flash: Smarter, Faster, More Efficient",
      "source_date": "2026-09-10",
      "where": "benchmark table image (v4.1_260910_benchmark_table_en), row 'SEC-Bench Pro', column 'DeepSeek V4.1-Flash'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "Claude Code harness; max reasoning effort",
      "caveat": "",
      "human_verified": false,
      "human_verified_on": null,
      "check": "corroborated",
      "corroborated": true,
      "corroborated_on": "2026-10-30",
      "corroborated_by": [
        {
          "url": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash",
          "name": "DeepSeek-V4.1-Flash card",
          "title": "deepseek-ai/DeepSeek-V4.1-Flash model card",
          "lab": "deepseek",
          "date": "2026-09-10",
          "printed": "62.8",
          "where": "README, 'Comparison with frontier models (Max reasoning effort)' table, row 'SEC-Bench Pro (Pass@1)', column 'DS-V4.1-Flash'"
        },
        {
          "url": "https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-4-carbon/",
          "name": "Gemini 4 Carbon",
          "title": "Gemini 4 Carbon: our next frontier model",
          "lab": "google",
          "date": "2026-10-28",
          "printed": "62.8%",
          "where": "SEC-bench Pro bar chart (Pass@1), value label on bar 'DeepSeek-V4.1 -Flash'"
        },
        {
          "url": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash/blob/main/DeepSeek_V41_Tech_Report.pdf",
          "name": "DeepSeek-V4.1-Flash card",
          "title": "DeepSeek-V4.1-Flash: Pushing the Limits of KV Cache Compression",
          "lab": "deepseek",
          "date": "2026-09-10",
          "printed": "62.8",
          "where": "Table 3, row 'SEC-Bench Pro (Pass@1)', column 'DS-V4.1-Flash' (Max)"
        }
      ],
      "source_image": "https://frontierlog.ai/sources/19d3b83b1b9e5f5b.png",
      "source_image_original": "https://api-docs.deepseek.com/img/v4.1_260910_benchmark_table_en.png"
    },
    {
      "id": "gemini-4-argon:sec-bench-pro-260505:default:1",
      "model": "gemini-4-argon",
      "model_name": "Gemini 4 Argon",
      "model_date": "2026-09-30",
      "event": "https://frontierlog.ai/event/2026-09-30-gemini-4-argon",
      "benchmark": "sec-bench-pro-260505",
      "benchmark_name": "SEC-bench Pro 260505",
      "setting": "default",
      "setting_name": "",
      "score": 83.4,
      "unit": "%",
      "printed": "83.4%",
      "reported_by": "google",
      "run_by": null,
      "label": "Gemini 4 Argon",
      "source": "https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-4-carbon/",
      "source_title": "Gemini 4 Carbon: our next frontier model",
      "source_date": "2026-10-28",
      "where": "benchmark table (image), Cybersecurity group, row 'SEC-bench Pro', column 'Gemini 4 Argon'",
      "own": false,
      "primary": true,
      "parallel": false,
      "notes": "183-task JavaScript engine split (May 2026); pass@1, mean of 3 attempts; Antigravity harness; LLM judge for crashes in both builds",
      "caveat": "",
      "human_verified": false,
      "human_verified_on": null,
      "check": "corroborated",
      "corroborated": true,
      "corroborated_on": "2026-10-30",
      "corroborated_by": [
        {
          "url": "https://storage.googleapis.com/deepmind-media/gemini/gemini_4_carbon_model_evaluation.pdf",
          "name": "Gemini 4 Carbon card",
          "title": "Gemini 4 Carbon Model evaluation: approach, methodology & results",
          "lab": "google",
          "date": "2026-10-28",
          "printed": "83.4%",
          "where": "page 5 'Results' table (raster image), row 'SEC-bench Pro' (Cybersecurity), column 'Gemini 4 Argon'"
        }
      ],
      "source_image": "https://frontierlog.ai/sources/bc9b1176d783fc3d.png",
      "source_image_original": "https://storage.googleapis.com/gweb-uniblog-publish-prod/images/gemini_4_carbon__evals-table.original.png"
    },
    {
      "id": "gemini-4-carbon:sec-bench-pro-260505:default:1",
      "model": "gemini-4-carbon",
      "model_name": "Gemini 4 Carbon",
      "model_date": "2026-10-28",
      "event": "https://frontierlog.ai/event/2026-10-28-gemini-4-carbon",
      "benchmark": "sec-bench-pro-260505",
      "benchmark_name": "SEC-bench Pro 260505",
      "setting": "default",
      "setting_name": "",
      "score": 88.2,
      "unit": "%",
      "printed": "88.2%",
      "reported_by": "google",
      "run_by": null,
      "label": "Gemini 4 Carbon",
      "source": "https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-4-carbon/",
      "source_title": "Gemini 4 Carbon: our next frontier model",
      "source_date": "2026-10-28",
      "where": "benchmark table (image), Cybersecurity group, row 'SEC-bench Pro', column 'Gemini 4 Carbon'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "183-task JavaScript engine split (May 2026); pass@1, mean of 3 attempts; Antigravity harness; LLM judge for crashes in both builds",
      "caveat": "",
      "human_verified": false,
      "human_verified_on": null,
      "check": "corroborated",
      "corroborated": true,
      "corroborated_on": "2026-10-30",
      "corroborated_by": [
        {
          "url": "https://storage.googleapis.com/deepmind-media/gemini/gemini_4_carbon_model_evaluation.pdf",
          "name": "Gemini 4 Carbon card",
          "title": "Gemini 4 Carbon Model evaluation: approach, methodology & results",
          "lab": "google",
          "date": "2026-10-28",
          "printed": "88.2%",
          "where": "page 5 'Results' table (raster image), row 'SEC-bench Pro' (Cybersecurity), column 'Gemini 4 Carbon'"
        }
      ],
      "source_image": "https://frontierlog.ai/sources/bc9b1176d783fc3d.png",
      "source_image_original": "https://storage.googleapis.com/gweb-uniblog-publish-prod/images/gemini_4_carbon__evals-table.original.png"
    }
  ]
}