{
  "api_version": "1",
  "updated": "2026-10-26",
  "latest_event": "2026-10-24",
  "rights": "© Frontier Log, all rights reserved.",
  "model": {
    "slug": "deepseek-v3-1",
    "name": "DeepSeek-V3.1",
    "lab": "deepseek",
    "date": "2025-08-21",
    "source": "https://api-docs.deepseek.com/news/news250821/",
    "source_title": "DeepSeek-V3.1 Release",
    "flagship": false,
    "event": "https://frontierlog.ai/event/2025-08-21-deepseek-v3-1",
    "releases": [],
    "variant_of": null,
    "made_by": "DeepSeek",
    "url": "https://frontierlog.ai/model/deepseek-v3-1",
    "results": [
      {
        "id": "deepseek-v3-1:swe-bench-verified:default:1",
        "model": "deepseek-v3-1",
        "model_name": "DeepSeek-V3.1",
        "model_date": "2025-08-21",
        "event": "https://frontierlog.ai/event/2025-08-21-deepseek-v3-1",
        "benchmark": "swe-bench-verified",
        "benchmark_name": "SWE-bench Verified",
        "setting": "default",
        "setting_name": "",
        "score": 66.0,
        "unit": "%",
        "printed": "66.0",
        "reported_by": "deepseek",
        "run_by": null,
        "label": "DeepSeek-V3.1",
        "source": "https://api-docs.deepseek.com/news/news250821/",
        "source_title": "DeepSeek-V3.1 Release",
        "source_date": "2025-08-21",
        "where": "first table image (agent coding, under 'Tools & Agents Upgrades'), row 'SWE-bench Verified', column 'DeepSeek-V3.1' (shaded)",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "non-thinking mode, as an agent",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-19",
        "check": "human",
        "corroborated": true,
        "corroborated_on": "2026-10-19",
        "corroborated_by": [
          {
            "url": "https://qwen.ai/blog?id=qwen3-max",
            "name": "Qwen3-Max",
            "title": "Qwen3-Max: Just Scale it",
            "lab": "alibaba",
            "date": "2025-09-24",
            "printed": "66.0",
            "where": "Qwen3-Max-Instruct bar chart (Qwen3-Max-Instruct.png), group SWE-Bench (Verified), value label on the DeepSeek-V3.1 Non-thinking bar"
          }
        ],
        "source_image": "https://frontierlog.ai/sources/7492c42eb0e15de5.webp",
        "source_image_original": "https://cdn.deepseek.com/api-docs/v3.1_benchmark_1.webp"
      },
      {
        "id": "deepseek-v3-1:terminal-bench-1-0:default:1",
        "model": "deepseek-v3-1",
        "model_name": "DeepSeek-V3.1",
        "model_date": "2025-08-21",
        "event": "https://frontierlog.ai/event/2025-08-21-deepseek-v3-1",
        "benchmark": "terminal-bench-1-0",
        "benchmark_name": "Terminal-Bench 1.0",
        "setting": "default",
        "setting_name": "",
        "score": 31.3,
        "unit": "%",
        "printed": "31.3",
        "reported_by": "deepseek",
        "run_by": null,
        "label": "DeepSeek-V3.1",
        "source": "https://api-docs.deepseek.com/news/news250821/",
        "source_title": "DeepSeek-V3.1 Release",
        "source_date": "2025-08-21",
        "where": "first table image (agent coding), row 'Terminal-Bench', column 'DeepSeek-V3.1' (shaded)",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-19",
        "check": "human",
        "corroborated": false,
        "corroborated_on": null,
        "corroborated_by": [],
        "source_image": "https://frontierlog.ai/sources/7492c42eb0e15de5.webp",
        "source_image_original": "https://cdn.deepseek.com/api-docs/v3.1_benchmark_1.webp"
      },
      {
        "id": "deepseek-v3-1:browsecomp:default:1",
        "model": "deepseek-v3-1",
        "model_name": "DeepSeek-V3.1",
        "model_date": "2025-08-21",
        "event": "https://frontierlog.ai/event/2025-08-21-deepseek-v3-1",
        "benchmark": "browsecomp",
        "benchmark_name": "BrowseComp",
        "setting": "default",
        "setting_name": "",
        "score": 30.0,
        "unit": "%",
        "printed": "30.0",
        "reported_by": "deepseek",
        "run_by": null,
        "label": "DeepSeek-V3.1",
        "source": "https://api-docs.deepseek.com/news/news250821/",
        "source_title": "DeepSeek-V3.1 Release",
        "source_date": "2025-08-21",
        "where": "second table image (search agent), row 'Browsecomp', column 'DeepSeek-V3.1' (shaded)",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "search agent",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-19",
        "check": "human",
        "corroborated": false,
        "corroborated_on": null,
        "corroborated_by": [],
        "source_image": "https://frontierlog.ai/sources/c164a27411cd68bc.webp",
        "source_image_original": "https://cdn.deepseek.com/api-docs/v3.1_benchmark_2.webp"
      },
      {
        "id": "deepseek-v3-1:hle:with-tools:1",
        "model": "deepseek-v3-1",
        "model_name": "DeepSeek-V3.1",
        "model_date": "2025-08-21",
        "event": "https://frontierlog.ai/event/2025-08-21-deepseek-v3-1",
        "benchmark": "hle",
        "benchmark_name": "Humanity's Last Exam",
        "setting": "with-tools",
        "setting_name": "With tools",
        "score": 29.8,
        "unit": "%",
        "printed": "29.8",
        "reported_by": "deepseek",
        "run_by": null,
        "label": "DeepSeek-V3.1",
        "source": "https://api-docs.deepseek.com/news/news250821/",
        "source_title": "DeepSeek-V3.1 Release",
        "source_date": "2025-08-21",
        "where": "second table image (search agent), row 'HLE', column 'DeepSeek-V3.1' (shaded)",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "search agent",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-19",
        "check": "human",
        "corroborated": false,
        "corroborated_on": null,
        "corroborated_by": [],
        "source_image": "https://frontierlog.ai/sources/c164a27411cd68bc.webp",
        "source_image_original": "https://cdn.deepseek.com/api-docs/v3.1_benchmark_2.webp"
      },
      {
        "id": "deepseek-v3-1:aime-2025:no-tools:1",
        "model": "deepseek-v3-1",
        "model_name": "DeepSeek-V3.1",
        "model_date": "2025-08-21",
        "event": "https://frontierlog.ai/event/2025-08-21-deepseek-v3-1",
        "benchmark": "aime-2025",
        "benchmark_name": "AIME 2025",
        "setting": "no-tools",
        "setting_name": "No tools",
        "score": 88.4,
        "unit": "%",
        "printed": "88.4%",
        "reported_by": "deepseek",
        "run_by": null,
        "label": "V3.1-Think",
        "source": "https://api-docs.deepseek.com/news/news250821/",
        "source_title": "DeepSeek-V3.1 Release",
        "source_date": "2025-08-21",
        "where": "third image (thinking efficiency bar chart), group 'AIME 2025', hatched bar 'V3.1-Think', label above the bar (output tokens, then the score in parentheses)",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "thinking mode; 15,889 output tokens on average",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-19",
        "check": "human",
        "corroborated": false,
        "corroborated_on": null,
        "corroborated_by": [],
        "source_image": "https://frontierlog.ai/sources/46ed66f55ee72932.webp",
        "source_image_original": "https://cdn.deepseek.com/api-docs/v3.1_benchmark_3.webp"
      },
      {
        "id": "deepseek-v3-1:gpqa-diamond:default:1",
        "model": "deepseek-v3-1",
        "model_name": "DeepSeek-V3.1",
        "model_date": "2025-08-21",
        "event": "https://frontierlog.ai/event/2025-08-21-deepseek-v3-1",
        "benchmark": "gpqa-diamond",
        "benchmark_name": "GPQA Diamond",
        "setting": "default",
        "setting_name": "",
        "score": 80.1,
        "unit": "%",
        "printed": "80.1%",
        "reported_by": "deepseek",
        "run_by": null,
        "label": "V3.1-Think",
        "source": "https://api-docs.deepseek.com/news/news250821/",
        "source_title": "DeepSeek-V3.1 Release",
        "source_date": "2025-08-21",
        "where": "third image (thinking efficiency bar chart), group 'GPQA Diamond', hatched bar 'V3.1-Think', label above the bar (output tokens, then the score in parentheses)",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "thinking mode; 4,122 output tokens on average",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-19",
        "check": "human",
        "corroborated": false,
        "corroborated_on": null,
        "corroborated_by": [],
        "source_image": "https://frontierlog.ai/sources/46ed66f55ee72932.webp",
        "source_image_original": "https://cdn.deepseek.com/api-docs/v3.1_benchmark_3.webp"
      },
      {
        "id": "deepseek-v3-1:livecodebench:default:1",
        "model": "deepseek-v3-1",
        "model_name": "DeepSeek-V3.1",
        "model_date": "2025-08-21",
        "event": "https://frontierlog.ai/event/2025-08-21-deepseek-v3-1",
        "benchmark": "livecodebench",
        "benchmark_name": "LiveCodeBench",
        "setting": "default",
        "setting_name": "",
        "score": 74.8,
        "unit": "%",
        "printed": "74.8%",
        "reported_by": "deepseek",
        "run_by": null,
        "label": "V3.1-Think",
        "source": "https://api-docs.deepseek.com/news/news250821/",
        "source_title": "DeepSeek-V3.1 Release",
        "source_date": "2025-08-21",
        "where": "third image (thinking efficiency bar chart), group 'LiveCodeBench', hatched bar 'V3.1-Think', label above the bar (output tokens, then the score in parentheses)",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "thinking mode; 13,977 output tokens on average; problem window not stated",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-19",
        "check": "human",
        "corroborated": false,
        "corroborated_on": null,
        "corroborated_by": [],
        "source_image": "https://frontierlog.ai/sources/46ed66f55ee72932.webp",
        "source_image_original": "https://cdn.deepseek.com/api-docs/v3.1_benchmark_3.webp"
      },
      {
        "id": "deepseek-v3-1:hle:no-tools:1",
        "model": "deepseek-v3-1",
        "model_name": "DeepSeek-V3.1",
        "model_date": "2025-08-21",
        "event": "https://frontierlog.ai/event/2025-08-21-deepseek-v3-1",
        "benchmark": "hle",
        "benchmark_name": "Humanity's Last Exam",
        "setting": "no-tools",
        "setting_name": "No tools",
        "score": 15.9,
        "unit": "%",
        "printed": "15.9",
        "reported_by": "deepseek",
        "run_by": null,
        "label": "DeepSeek-V3.1",
        "source": "https://api-docs.deepseek.com/news/news250922/",
        "source_title": "DeepSeek-V3.1-Terminus",
        "source_date": "2025-09-22",
        "where": "later report: DeepSeek-V3.1-Terminus post, table image, section 'reasoning mode w/o tool use', row \"Humanity's Last Exam\", column 'DeepSeek-V3.1'",
        "own": false,
        "primary": true,
        "parallel": false,
        "notes": "reasoning mode; full or text-only set not stated",
        "caveat": "",
        "human_verified": false,
        "human_verified_on": null,
        "check": "corroborated",
        "corroborated": true,
        "corroborated_on": "2026-10-19",
        "corroborated_by": [
          {
            "url": "https://huggingface.co/deepseek-ai/DeepSeek-V3.1",
            "name": "DeepSeek-V3.1 card",
            "title": "deepseek-ai/DeepSeek-V3.1 model card",
            "lab": "deepseek",
            "date": "2025-08-21",
            "printed": "15.9",
            "where": "Evaluation table, General, row 'Humanity's Last Exam (Pass@1)', column 'DeepSeek V3.1-Thinking' (note: 'HLE is evaluated with the text-only subset')"
          },
          {
            "url": "https://huggingface.co/deepseek-ai/DeepSeek-V3.1-Terminus",
            "name": "DeepSeek-V3.1-Terminus card",
            "title": "deepseek-ai/DeepSeek-V3.1-Terminus model card",
            "lab": "deepseek",
            "date": "2025-09-22",
            "printed": "15.9",
            "where": "Table, 'Reasoning Mode w/o Tool Use', row 'Humanity's Last Exam', column 'DeepSeek-V3.1'"
          }
        ],
        "source_image": "https://frontierlog.ai/sources/93e083b75de124e9.png",
        "source_image_original": "https://api-docs.deepseek.com/zh-cn/img/deepseek_v3.1_terminus_benchmark_en.png"
      }
    ]
  }
}