GPT-5.4

{
  "id": "openai/gpt-5.4",
  "name": "GPT-5.4",
  "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost",
  "family": "gpt",
  "last_updated": "2026-03-05",
  "knowledge": "2025-08-31",
  "attachment": true,
  "reasoning": true,
  "tool_call": true,
  "structured_output": true,
  "temperature": true,
  "modalities": {
    "input": [
      "text",
      "image",
      "pdf"
    ],
    "output": [
      "text"
    ]
  },
  "limit": {
    "context": 1050000,
    "input": 922000,
    "output": 128000
  },
  "benchmarks": [
    {
      "name": "SWE-Bench Pro",
      "score": 59.1,
      "source": "https://labs.scale.com/leaderboard/swe_bench_pro_public",
      "metric": "resolve rate",
      "dataset": "public"
    },
    {
      "name": "SWE-Atlas Codebase QnA",
      "score": 40.8,
      "source": "https://labs.scale.com/leaderboard/sweatlas-qna",
      "metric": "score",
      "harness": "Codex"
    },
    {
      "name": "SWE-Atlas Codebase QnA",
      "score": 36.3,
      "source": "https://labs.scale.com/leaderboard/sweatlas-qna",
      "metric": "score",
      "harness": "Mini-SWE-Agent"
    },
    {
      "name": "SWE-Atlas Refactoring",
      "score": 44.29,
      "source": "https://labs.scale.com/leaderboard/sweatlas-refactoring",
      "metric": "score",
      "harness": "Codex"
    },
    {
      "name": "SWE-Atlas Test Writing",
      "score": 44.36,
      "source": "https://labs.scale.com/leaderboard/sweatlas-tw",
      "metric": "score",
      "harness": "Codex CLI"
    },
    {
      "name": "SWE-Atlas Test Writing",
      "score": 40,
      "source": "https://labs.scale.com/leaderboard/sweatlas-tw",
      "metric": "score",
      "harness": "Mini-SWE-Agent"
    },
    {
      "name": "Artificial Analysis Coding Agent Index",
      "score": 53.6,
      "source": "https://artificialanalysis.ai/agents/coding-agents",
      "metric": "average pass@1",
      "harness": "Codex",
      "variant": "medium"
    },
    {
      "name": "SWE-Atlas Codebase QnA",
      "score": 72.4,
      "source": "https://artificialanalysis.ai/agents/coding-agents",
      "metric": "pass@1",
      "harness": "Codex",
      "variant": "medium"
    },
    {
      "name": "SWE-Bench Pro",
      "score": 18.4,
      "source": "https://artificialanalysis.ai/agents/coding-agents",
      "metric": "pass@1",
      "harness": "Codex",
      "variant": "medium",
      "dataset": "hard-aa"
    },
    {
      "name": "Terminal-Bench",
      "score": 69.8,
      "source": "https://artificialanalysis.ai/agents/coding-agents",
      "metric": "pass@1",
      "harness": "Codex",
      "variant": "medium",
      "version": "2.1"
    },
    {
      "name": "Artificial Analysis Coding Agent Index",
      "score": 52.2,
      "source": "https://artificialanalysis.ai/agents/coding-agents",
      "metric": "average pass@1",
      "harness": "Cursor CLI",
      "variant": "medium"
    },
    {
      "name": "SWE-Atlas Codebase QnA",
      "score": 72.9,
      "source": "https://artificialanalysis.ai/agents/coding-agents",
      "metric": "pass@1",
      "harness": "Cursor CLI",
      "variant": "medium"
    },
    {
      "name": "SWE-Bench Pro",
      "score": 18.9,
      "source": "https://artificialanalysis.ai/agents/coding-agents",
      "metric": "pass@1",
      "harness": "Cursor CLI",
      "variant": "medium",
      "dataset": "hard-aa"
    },
    {
      "name": "Terminal-Bench",
      "score": 64.7,
      "source": "https://artificialanalysis.ai/agents/coding-agents",
      "metric": "pass@1",
      "harness": "Cursor CLI",
      "variant": "medium",
      "version": "2.1"
    },
    {
      "name": "Terminal-Bench",
      "score": 75.1,
      "source": "https://openai.com/index/introducing-gpt-5-5/",
      "metric": "success rate",
      "version": "2.0",
      "date": "2026-04-23"
    },
    {
      "name": "GPQA Diamond",
      "score": 92.8,
      "source": "https://openai.com/index/introducing-gpt-5-5/",
      "metric": "accuracy",
      "date": "2026-04-23"
    },
    {
      "name": "Humanity's Last Exam",
      "score": 39.8,
      "source": "https://openai.com/index/introducing-gpt-5-5/",
      "metric": "accuracy",
      "variant": "no tools",
      "date": "2026-04-23"
    },
    {
      "name": "Humanity's Last Exam",
      "score": 52.1,
      "source": "https://openai.com/index/introducing-gpt-5-5/",
      "metric": "accuracy",
      "variant": "with tools",
      "date": "2026-04-23"
    },
    {
      "name": "OSWorld-Verified",
      "score": 75,
      "source": "https://openai.com/index/introducing-gpt-5-5/",
      "metric": "success rate",
      "date": "2026-04-23"
    },
    {
      "name": "BrowseComp",
      "score": 82.7,
      "source": "https://openai.com/index/introducing-gpt-5-5/",
      "metric": "accuracy",
      "date": "2026-04-23"
    },
    {
      "name": "GDPval",
      "score": 83,
      "source": "https://openai.com/index/introducing-gpt-5-5/",
      "metric": "wins or ties",
      "date": "2026-04-23"
    },
    {
      "name": "ARC-AGI-2",
      "score": 73.3,
      "source": "https://openai.com/index/introducing-gpt-5-5/",
      "metric": "accuracy",
      "variant": "Verified",
      "date": "2026-04-23"
    },
    {
      "name": "FrontierMath",
      "score": 47.6,
      "source": "https://openai.com/index/introducing-gpt-5-5/",
      "metric": "accuracy",
      "dataset": "Tier 1-3",
      "date": "2026-04-23"
    },
    {
      "name": "FrontierMath",
      "score": 27.1,
      "source": "https://openai.com/index/introducing-gpt-5-5/",
      "metric": "accuracy",
      "dataset": "Tier 4",
      "date": "2026-04-23"
    },
    {
      "name": "MMMU Pro",
      "score": 81.2,
      "source": "https://openai.com/index/introducing-gpt-5-5/",
      "metric": "accuracy",
      "variant": "no tools",
      "date": "2026-04-23"
    }
  ],
  "release_date": "2026-03-05",
  "open_weights": false
}