DeepSeek V4.1 Flash

{
  "id": "deepseek/deepseek-v4.1-flash",
  "name": "DeepSeek V4.1 Flash",
  "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding",
  "family": "deepseek-flash",
  "last_updated": "2026-09-10",
  "knowledge": "2025-05",
  "attachment": true,
  "reasoning": true,
  "tool_call": true,
  "structured_output": true,
  "temperature": true,
  "modalities": {
    "input": [
      "text",
      "image"
    ],
    "output": [
      "text"
    ]
  },
  "limit": {
    "context": 1000000,
    "output": 384000
  },
  "benchmarks": [
    {
      "name": "GPQA Diamond",
      "score": 90.9,
      "source": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash",
      "metric": "pass@1",
      "variant": "reasoning_effort=100"
    },
    {
      "name": "MathArena Apex",
      "score": 65.6,
      "source": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash",
      "metric": "pass@1",
      "variant": "reasoning_effort=100"
    },
    {
      "name": "Humanity's Last Exam",
      "score": 36.8,
      "source": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash",
      "metric": "pass@1",
      "variant": "reasoning_effort=100",
      "dataset": "full set"
    },
    {
      "name": "Humanity's Last Exam",
      "score": 39.1,
      "source": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash",
      "metric": "pass@1",
      "variant": "reasoning_effort=100",
      "dataset": "text-only subset"
    },
    {
      "name": "Codeforces",
      "score": 3471,
      "source": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash",
      "metric": "rating",
      "variant": "reasoning_effort=100"
    },
    {
      "name": "Terminal-Bench",
      "score": 90.6,
      "source": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash",
      "metric": "pass@1",
      "harness": "DeepSeek Harness minimal mode",
      "variant": "reasoning_effort=100",
      "version": "2.1"
    },
    {
      "name": "Terminal-Bench",
      "score": 30,
      "source": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash",
      "metric": "pass@1",
      "harness": "DeepSeek Harness minimal mode",
      "variant": "reasoning_effort=100",
      "version": "3.0"
    },
    {
      "name": "Terminal-Bench",
      "score": 31.2,
      "source": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash",
      "metric": "pass@1",
      "harness": "DeepSeek Harness minimal mode",
      "variant": "reasoning_effort=100",
      "version": "4.0"
    },
    {
      "name": "DeepSWE",
      "score": 74.2,
      "source": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash",
      "metric": "resolved",
      "harness": "mini-swe-agent",
      "variant": "reasoning_effort=100",
      "version": "1.1"
    },
    {
      "name": "ProgramBench",
      "score": 20.3,
      "source": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash",
      "metric": "almost@1",
      "harness": "DeepSeek Harness minimal mode",
      "variant": "reasoning_effort=100"
    },
    {
      "name": "NL2Repo",
      "score": 64,
      "source": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash",
      "metric": "score",
      "harness": "DeepSeek Harness minimal mode",
      "variant": "reasoning_effort=100"
    },
    {
      "name": "CyberGym",
      "score": 88.1,
      "source": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash",
      "metric": "pass@1",
      "variant": "reasoning_effort=100"
    },
    {
      "name": "ExploitGym",
      "score": 15.3,
      "source": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash",
      "metric": "pass@1",
      "variant": "reasoning_effort=100"
    },
    {
      "name": "AutomationBench",
      "score": 54.8,
      "source": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash",
      "metric": "pass@1",
      "variant": "reasoning_effort=100"
    },
    {
      "name": "Agents' Last Exam",
      "score": 31.8,
      "source": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash",
      "metric": "pass@1",
      "variant": "reasoning_effort=100"
    },
    {
      "name": "SEC-Bench Pro",
      "score": 62.8,
      "source": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash",
      "metric": "pass@1",
      "harness": "Claude Code",
      "variant": "reasoning_effort=100"
    },
    {
      "name": "Humanity's Last Exam",
      "score": 63.9,
      "source": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash",
      "metric": "pass@1",
      "variant": "reasoning_effort=100; with tools"
    },
    {
      "name": "Chartography",
      "score": 78.9,
      "source": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash",
      "metric": "pass@1",
      "harness": "Claude Code",
      "variant": "reasoning_effort=100; with tools"
    },
    {
      "name": "BabyVision",
      "score": 89.6,
      "source": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash",
      "metric": "pass@1",
      "harness": "Claude Code",
      "variant": "reasoning_effort=100; with tools"
    },
    {
      "name": "ZeroBench",
      "score": 49,
      "source": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash",
      "metric": "pass@5",
      "harness": "Claude Code",
      "variant": "reasoning_effort=100; with tools",
      "dataset": "main"
    },
    {
      "name": "DeepSWE",
      "score": 69.8,
      "source": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash",
      "metric": "resolved",
      "harness": "Claude Code",
      "variant": "reasoning_effort=100",
      "version": "1.1"
    },
    {
      "name": "Terminal-Bench",
      "score": 88,
      "source": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash",
      "metric": "pass@1",
      "harness": "Claude Code",
      "variant": "reasoning_effort=100",
      "version": "2.1"
    },
    {
      "name": "DeepSWE",
      "score": 65.6,
      "source": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash",
      "metric": "resolved",
      "harness": "Codex",
      "variant": "reasoning_effort=100",
      "version": "1.1"
    },
    {
      "name": "Terminal-Bench",
      "score": 84.1,
      "source": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash",
      "metric": "pass@1",
      "harness": "Codex",
      "variant": "reasoning_effort=100",
      "version": "2.1"
    },
    {
      "name": "DeepSWE",
      "score": 65.5,
      "source": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash",
      "metric": "resolved",
      "harness": "OpenCode",
      "variant": "reasoning_effort=100",
      "version": "1.1"
    },
    {
      "name": "Terminal-Bench",
      "score": 85,
      "source": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash",
      "metric": "pass@1",
      "harness": "OpenCode",
      "variant": "reasoning_effort=100",
      "version": "2.1"
    },
    {
      "name": "DeepSWE",
      "score": 66.2,
      "source": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash",
      "metric": "resolved",
      "harness": "Pi",
      "variant": "reasoning_effort=100",
      "version": "1.1"
    },
    {
      "name": "Terminal-Bench",
      "score": 86.1,
      "source": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash",
      "metric": "pass@1",
      "harness": "Pi",
      "variant": "reasoning_effort=100",
      "version": "2.1"
    },
    {
      "name": "Terminal-Bench",
      "score": 90.3,
      "source": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash",
      "metric": "pass@1",
      "harness": "mini-swe-agent",
      "variant": "reasoning_effort=100",
      "version": "2.1"
    },
    {
      "name": "DeepSWE",
      "score": 72.6,
      "source": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash",
      "metric": "resolved",
      "harness": "DeepSeek Harness minimal mode",
      "variant": "reasoning_effort=100",
      "version": "1.1"
    },
    {
      "name": "DeepSWE",
      "score": 70.5,
      "source": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash",
      "metric": "resolved",
      "harness": "DeepSeek Harness standard mode",
      "variant": "reasoning_effort=100",
      "version": "1.1"
    },
    {
      "name": "Terminal-Bench",
      "score": 85.8,
      "source": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash",
      "metric": "pass@1",
      "harness": "DeepSeek Harness standard mode",
      "variant": "reasoning_effort=100",
      "version": "2.1"
    },
    {
      "name": "DeepSWE",
      "score": 67.6,
      "source": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash",
      "metric": "resolved",
      "harness": "DeepSeek Harness PTC mode",
      "variant": "reasoning_effort=100",
      "version": "1.1"
    },
    {
      "name": "Terminal-Bench",
      "score": 85.8,
      "source": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash",
      "metric": "pass@1",
      "harness": "DeepSeek Harness PTC mode",
      "variant": "reasoning_effort=100",
      "version": "2.1"
    }
  ],
  "release_date": "2026-09-10",
  "open_weights": true,
  "license": "MIT",
  "weights": [
    {
      "label": "Hugging Face",
      "url": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash"
    }
  ]
}