{
  "api_version": "1",
  "updated": "2026-10-26",
  "latest_event": "2026-10-24",
  "rights": "© Frontier Log, all rights reserved.",
  "model": {
    "slug": "deepseek-v4-pro",
    "name": "DeepSeek-V4-Pro",
    "lab": "deepseek",
    "date": "2026-04-24",
    "flagship": false,
    "event": "https://frontierlog.ai/event/2026-04-24-deepseek-v4-preview",
    "releases": [],
    "variant_of": null,
    "made_by": "DeepSeek",
    "url": "https://frontierlog.ai/model/deepseek-v4-pro",
    "results": [
      {
        "id": "deepseek-v4-pro:swe-bench-verified:default:1",
        "model": "deepseek-v4-pro",
        "model_name": "DeepSeek-V4-Pro",
        "model_date": "2026-04-24",
        "event": "https://frontierlog.ai/event/2026-04-24-deepseek-v4-preview",
        "benchmark": "swe-bench-verified",
        "benchmark_name": "SWE-bench Verified",
        "setting": "default",
        "setting_name": "",
        "score": 80.6,
        "unit": "%",
        "printed": "80.6",
        "reported_by": "deepseek",
        "run_by": null,
        "label": "DS-V4-Pro Max",
        "source": "https://api-docs.deepseek.com/news/news260424/",
        "source_title": "DeepSeek V4 Preview Release",
        "source_date": "2026-04-24",
        "where": "benchmark table image, row 'SWE Verified (Resolved)', column 'DS-V4-Pro Max'",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "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://www.minimax.io/blog/minimax-m3",
            "name": "MiniMax M3",
            "title": "MiniMax M3: Frontier Coding, 1M Context, Native Multimodality — All in One Model",
            "lab": "minimax",
            "date": "2026-06-01",
            "printed": "80.6",
            "where": "Full benchmark table image; row 'SWE-Bench Verified'; column 'DeepSeek V4 Pro'"
          },
          {
            "url": "https://thinkingmachines.ai/news/introducing-inkling/",
            "name": "Inkling",
            "title": "Inkling: Our Open-Weights Model",
            "lab": "thinking-machines",
            "date": "2026-07-15",
            "printed": "80.6%",
            "where": "'Benchmarking Inkling' main table, row 'SWEBench Verified* (Inkling bash-only harness; external models self-reported)', column 'DeepSeek V4 Pro'"
          }
        ],
        "source_image": "https://frontierlog.ai/sources/9226203de8e24b33.png",
        "source_image_original": "https://api-docs.deepseek.com/img/v4-benchmark-2.png"
      },
      {
        "id": "deepseek-v4-pro:gpqa-diamond:default:1",
        "model": "deepseek-v4-pro",
        "model_name": "DeepSeek-V4-Pro",
        "model_date": "2026-04-24",
        "event": "https://frontierlog.ai/event/2026-04-24-deepseek-v4-preview",
        "benchmark": "gpqa-diamond",
        "benchmark_name": "GPQA Diamond",
        "setting": "default",
        "setting_name": "",
        "score": 90.1,
        "unit": "%",
        "printed": "90.1",
        "reported_by": "deepseek",
        "run_by": null,
        "label": "DS-V4-Pro Max",
        "source": "https://api-docs.deepseek.com/news/news260424/",
        "source_title": "DeepSeek V4 Preview Release",
        "source_date": "2026-04-24",
        "where": "benchmark table image, row 'GPQA Diamond (Pass@1)', column 'DS-V4-Pro Max'",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "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://z.ai/blog/glm-5.2",
            "name": "GLM-5.2",
            "title": "GLM-5.2: Built for Long-Horizon Tasks",
            "lab": "zai",
            "date": "2026-06-16",
            "printed": "90.1",
            "where": "Full Benchmark Table, row 'GPQA-Diamond', column 'DeepSeek-V4-Pro'"
          }
        ],
        "source_image": "https://frontierlog.ai/sources/9226203de8e24b33.png",
        "source_image_original": "https://api-docs.deepseek.com/img/v4-benchmark-2.png"
      },
      {
        "id": "deepseek-v4-pro:browsecomp:default:1",
        "model": "deepseek-v4-pro",
        "model_name": "DeepSeek-V4-Pro",
        "model_date": "2026-04-24",
        "event": "https://frontierlog.ai/event/2026-04-24-deepseek-v4-preview",
        "benchmark": "browsecomp",
        "benchmark_name": "BrowseComp",
        "setting": "default",
        "setting_name": "",
        "score": 83.4,
        "unit": "%",
        "printed": "83.4",
        "reported_by": "deepseek",
        "run_by": null,
        "label": "DS-V4-Pro Max",
        "source": "https://api-docs.deepseek.com/news/news260424/",
        "source_title": "DeepSeek V4 Preview Release",
        "source_date": "2026-04-24",
        "where": "benchmark table image, row 'BrowseComp (Pass@1)' (Agentic), column 'DS-V4-Pro Max'",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "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://www.minimax.io/blog/minimax-m3",
            "name": "MiniMax M3",
            "title": "MiniMax M3: Frontier Coding, 1M Context, Native Multimodality — All in One Model",
            "lab": "minimax",
            "date": "2026-06-01",
            "printed": "83.4",
            "where": "Full benchmark table image; row 'BrowseComp'; column 'DeepSeek V4 Pro'"
          },
          {
            "url": "https://thinkingmachines.ai/news/introducing-inkling/",
            "name": "Inkling",
            "title": "Inkling: Our Open-Weights Model",
            "lab": "thinking-machines",
            "date": "2026-07-15",
            "printed": "83.4%",
            "where": "'Benchmarking Inkling' main table, row 'BrowseComp w/ ctx management', column 'DeepSeek V4 Pro'"
          }
        ],
        "source_image": "https://frontierlog.ai/sources/9226203de8e24b33.png",
        "source_image_original": "https://api-docs.deepseek.com/img/v4-benchmark-2.png"
      },
      {
        "id": "deepseek-v4-pro:swe-bench-pro:default:1",
        "model": "deepseek-v4-pro",
        "model_name": "DeepSeek-V4-Pro",
        "model_date": "2026-04-24",
        "event": "https://frontierlog.ai/event/2026-04-24-deepseek-v4-preview",
        "benchmark": "swe-bench-pro",
        "benchmark_name": "SWE-bench Pro",
        "setting": "default",
        "setting_name": "",
        "score": 55.4,
        "unit": "%",
        "printed": "55.4",
        "reported_by": "deepseek",
        "run_by": null,
        "label": "DS-V4-Pro Max",
        "source": "https://api-docs.deepseek.com/news/news260424/",
        "source_title": "DeepSeek V4 Preview Release",
        "source_date": "2026-04-24",
        "where": "benchmark table image, row 'SWE Pro (Resolved)' (Agentic), column 'DS-V4-Pro Max'",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "max reasoning effort",
        "caveat": "The source doesn't say whether this is the public or the commercial set",
        "human_verified": true,
        "human_verified_on": "2026-10-17",
        "check": "human",
        "corroborated": true,
        "corroborated_on": "2026-10-19",
        "corroborated_by": [
          {
            "url": "https://www.minimax.io/blog/minimax-m3",
            "name": "MiniMax M3",
            "title": "MiniMax M3: Frontier Coding, 1M Context, Native Multimodality — All in One Model",
            "lab": "minimax",
            "date": "2026-06-01",
            "printed": "55.4",
            "where": "Full benchmark table image; row 'SWE-Bench Pro'; column 'DeepSeek V4 Pro'"
          },
          {
            "url": "https://thinkingmachines.ai/news/introducing-inkling/",
            "name": "Inkling",
            "title": "Inkling: Our Open-Weights Model",
            "lab": "thinking-machines",
            "date": "2026-07-15",
            "printed": "55.4%",
            "where": "'Benchmarking Inkling' main table, row 'SWEBench Pro Public', column 'DeepSeek V4 Pro'"
          },
          {
            "url": "https://z.ai/blog/glm-5.2",
            "name": "GLM-5.2",
            "title": "GLM-5.2: Built for Long-Horizon Tasks",
            "lab": "zai",
            "date": "2026-06-16",
            "printed": "55.4",
            "where": "Full Benchmark Table, row 'SWE-bench Pro', column 'DeepSeek-V4-Pro'"
          }
        ],
        "source_image": "https://frontierlog.ai/sources/9226203de8e24b33.png",
        "source_image_original": "https://api-docs.deepseek.com/img/v4-benchmark-2.png"
      },
      {
        "id": "deepseek-v4-pro:terminal-bench-2-0:default:1",
        "model": "deepseek-v4-pro",
        "model_name": "DeepSeek-V4-Pro",
        "model_date": "2026-04-24",
        "event": "https://frontierlog.ai/event/2026-04-24-deepseek-v4-preview",
        "benchmark": "terminal-bench-2-0",
        "benchmark_name": "Terminal-Bench 2.0",
        "setting": "default",
        "setting_name": "",
        "score": 67.9,
        "unit": "%",
        "printed": "67.9",
        "reported_by": "deepseek",
        "run_by": null,
        "label": "DS-V4-Pro Max",
        "source": "https://api-docs.deepseek.com/news/news260424/",
        "source_title": "DeepSeek V4 Preview Release",
        "source_date": "2026-04-24",
        "where": "benchmark table image (v4-benchmark-2), section 'Agentic', row 'Terminal Bench 2.0 (Acc)', column 'DS-V4-Pro Max'",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "max reasoning effort",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-18",
        "check": "human",
        "corroborated": true,
        "corroborated_on": "2026-10-19",
        "corroborated_by": [
          {
            "url": "https://qwen.ai/blog?id=qwen3.7",
            "name": "Qwen3.7-Max",
            "title": "Qwen3.7: The Agent Frontier",
            "lab": "alibaba",
            "date": "2026-05-20",
            "printed": "67.9",
            "where": "Score chart image (Qwen3.7-Max-Score.png), panel Terminal-Bench 2.0 (Terminus-2), value label on the DS-V4-Pro Max bar"
          }
        ],
        "source_image": "https://frontierlog.ai/sources/9226203de8e24b33.png",
        "source_image_original": "https://api-docs.deepseek.com/img/v4-benchmark-2.png"
      },
      {
        "id": "deepseek-v4-pro:mmlu-pro:default:1",
        "model": "deepseek-v4-pro",
        "model_name": "DeepSeek-V4-Pro",
        "model_date": "2026-04-24",
        "event": "https://frontierlog.ai/event/2026-04-24-deepseek-v4-preview",
        "benchmark": "mmlu-pro",
        "benchmark_name": "MMLU-Pro",
        "setting": "default",
        "setting_name": "",
        "score": 87.5,
        "unit": "%",
        "printed": "87.5",
        "reported_by": "deepseek",
        "run_by": null,
        "label": "DS-V4-Pro Max",
        "source": "https://api-docs.deepseek.com/news/news260424/",
        "source_title": "DeepSeek V4 Preview Release",
        "source_date": "2026-04-24",
        "where": "benchmark table image, row 'MMLU-Pro (EM)', column 'DS-V4-Pro Max'",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "max reasoning effort",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-18",
        "check": "human",
        "corroborated": false,
        "corroborated_on": null,
        "corroborated_by": [],
        "source_image": "https://frontierlog.ai/sources/9226203de8e24b33.png",
        "source_image_original": "https://api-docs.deepseek.com/img/v4-benchmark-2.png"
      },
      {
        "id": "deepseek-v4-pro:livecodebench:default:1",
        "model": "deepseek-v4-pro",
        "model_name": "DeepSeek-V4-Pro",
        "model_date": "2026-04-24",
        "event": "https://frontierlog.ai/event/2026-04-24-deepseek-v4-preview",
        "benchmark": "livecodebench",
        "benchmark_name": "LiveCodeBench",
        "setting": "default",
        "setting_name": "",
        "score": 93.5,
        "unit": "%",
        "printed": "93.5",
        "reported_by": "deepseek",
        "run_by": null,
        "label": "DS-V4-Pro Max",
        "source": "https://api-docs.deepseek.com/news/news260424/",
        "source_title": "DeepSeek V4 Preview Release",
        "source_date": "2026-04-24",
        "where": "benchmark table image, row 'LiveCodeBench (Pass@1)', column 'DS-V4-Pro Max'",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "max reasoning effort; problem window not stated",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-18",
        "check": "human",
        "corroborated": false,
        "corroborated_on": null,
        "corroborated_by": [],
        "source_image": "https://frontierlog.ai/sources/9226203de8e24b33.png",
        "source_image_original": "https://api-docs.deepseek.com/img/v4-benchmark-2.png"
      },
      {
        "id": "deepseek-v4-pro:hle:no-tools:1",
        "model": "deepseek-v4-pro",
        "model_name": "DeepSeek-V4-Pro",
        "model_date": "2026-04-24",
        "event": "https://frontierlog.ai/event/2026-04-24-deepseek-v4-preview",
        "benchmark": "hle",
        "benchmark_name": "Humanity's Last Exam",
        "setting": "no-tools",
        "setting_name": "No tools",
        "score": 37.7,
        "unit": "%",
        "printed": "37.7",
        "reported_by": "deepseek",
        "run_by": null,
        "label": "DS-V4-Pro Max",
        "source": "https://api-docs.deepseek.com/news/news260424/",
        "source_title": "DeepSeek V4 Preview Release",
        "source_date": "2026-04-24",
        "where": "benchmark table image (v4-benchmark-2), section 'Knowledge & Reasoning', row 'HLE (Pass@1)', column 'DS-V4-Pro Max'",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "max reasoning effort; 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://api-docs.deepseek.com/news/news260813/",
            "name": "DeepSeek-V4-Pro GA",
            "title": "DeepSeek-V4-Pro GA Release",
            "lab": "deepseek",
            "date": "2026-08-13",
            "printed": "37.7",
            "where": "Benchmark table image, row 'HLE (wo/w tools)', column 'DeepSeek-V4-Pro-Preview', printed '37.7/48.2'"
          }
        ],
        "source_image": "https://frontierlog.ai/sources/9226203de8e24b33.png",
        "source_image_original": "https://api-docs.deepseek.com/img/v4-benchmark-2.png"
      },
      {
        "id": "deepseek-v4-pro:hle:with-tools:1",
        "model": "deepseek-v4-pro",
        "model_name": "DeepSeek-V4-Pro",
        "model_date": "2026-04-24",
        "event": "https://frontierlog.ai/event/2026-04-24-deepseek-v4-preview",
        "benchmark": "hle",
        "benchmark_name": "Humanity's Last Exam",
        "setting": "with-tools",
        "setting_name": "With tools",
        "score": 48.2,
        "unit": "%",
        "printed": "48.2",
        "reported_by": "deepseek",
        "run_by": null,
        "label": "DS-V4-Pro Max",
        "source": "https://api-docs.deepseek.com/news/news260424/",
        "source_title": "DeepSeek V4 Preview Release",
        "source_date": "2026-04-24",
        "where": "benchmark table image (v4-benchmark-2), section 'Agentic', row 'HLE w/tools (Pass@1)', column 'DS-V4-Pro Max'",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "max reasoning effort",
        "caveat": "",
        "human_verified": false,
        "human_verified_on": null,
        "check": "corroborated",
        "corroborated": true,
        "corroborated_on": "2026-10-19",
        "corroborated_by": [
          {
            "url": "https://api-docs.deepseek.com/news/news260813/",
            "name": "DeepSeek-V4-Pro GA",
            "title": "DeepSeek-V4-Pro GA Release",
            "lab": "deepseek",
            "date": "2026-08-13",
            "printed": "48.2",
            "where": "Benchmark table image, row 'HLE (wo/w tools)', column 'DeepSeek-V4-Pro-Preview', printed '37.7/48.2'"
          }
        ],
        "source_image": "https://frontierlog.ai/sources/9226203de8e24b33.png",
        "source_image_original": "https://api-docs.deepseek.com/img/v4-benchmark-2.png"
      },
      {
        "id": "deepseek-v4-pro:automationbench:default:1",
        "model": "deepseek-v4-pro",
        "model_name": "DeepSeek-V4-Pro",
        "model_date": "2026-04-24",
        "event": "https://frontierlog.ai/event/2026-04-24-deepseek-v4-preview",
        "benchmark": "automationbench",
        "benchmark_name": "AutomationBench",
        "setting": "default",
        "setting_name": "",
        "score": 12.8,
        "unit": "%",
        "printed": "12.8",
        "reported_by": "deepseek",
        "run_by": null,
        "label": "DeepSeek-V4-Pro-Preview",
        "source": "https://api-docs.deepseek.com/news/news260813/",
        "source_title": "DeepSeek-V4-Pro GA Release",
        "source_date": "2026-08-13",
        "where": "later report: DeepSeek-V4-Pro GA post, table image, row 'AutomationBench (Public)', column 'DeepSeek-V4-Pro-Preview'",
        "own": false,
        "primary": true,
        "parallel": false,
        "notes": "public set",
        "caveat": "",
        "human_verified": true,
        "human_verified_on": "2026-10-20",
        "check": "human",
        "corroborated": false,
        "corroborated_on": null,
        "corroborated_by": [],
        "source_image": "https://frontierlog.ai/sources/51578c812ffc7347.png",
        "source_image_original": "https://api-docs.deepseek.com/img/v4_260813_benchmark_table_en.png"
      }
    ]
  }
}