{
  "id": "gaia-benchmark",
  "type": "entity",
  "name": "GAIA benchmark",
  "summary": "A benchmark of 466 real-world questions for general AI assistants, designed so answers are unambiguous to grade but require reasoning, multi-modality, web browsing and tool use to reach. Its signature result is the human-AI gap: 92% for human respondents against 15% for GPT-4 with plugins at publication — questions conceptually simple for people, hard for tool-using models.",
  "locale": "en",
  "tags": [
    "gaia",
    "benchmarks",
    "assistants",
    "tool-use",
    "evaluation"
  ],
  "relations": [
    {
      "rel": "related",
      "target": "measuring-agent-reliability"
    },
    {
      "rel": "related",
      "target": "agentbench"
    }
  ],
  "questions": [
    "What is the GAIA benchmark?",
    "What does GAIA measure that other LLM benchmarks do not?"
  ],
  "claims": [
    {
      "id": "c1",
      "text": "GAIA comprises 466 real-world questions that jointly test reasoning, multi-modality handling, web browsing and general tool-use proficiency.",
      "sources": [
        "gaia-benchmark-paper"
      ],
      "confidence": 0.95
    },
    {
      "id": "c2",
      "text": "At publication, human respondents scored 92% on GAIA against 15% for GPT-4 equipped with plugins — the reverse of benchmarks where models beat humans on professional-exam material.",
      "sources": [
        "gaia-benchmark-paper"
      ],
      "confidence": 0.9
    }
  ],
  "takeaways": [],
  "faqs": [],
  "evidence_tier": "secondary",
  "evidence": {
    "level": "benchmark",
    "source_types": [
      "benchmark",
      "paper"
    ]
  },
  "moat_flag": false,
  "winning_edge": "States what GAIA's design actually selects for — gradeable answers reached only through tool chains — and reads its human-AI gap as an assistant-reliability signal, not a leaderboard curiosity.",
  "confidence": 0.85,
  "last_verified": "2026-08-08",
  "canonical_url": "https://agentreliability.dev/k/gaia-benchmark",
  "api_url": "https://agentreliability.dev/api/k/gaia-benchmark.json",
  "jsonld": {
    "@context": "https://schema.org",
    "name": "GAIA benchmark",
    "description": "A benchmark of 466 real-world questions for general AI assistants, designed so answers are unambiguous to grade but require reasoning, multi-modality, web browsing and tool use to reach. Its signature result is the human-AI gap: 92% for human respondents against 15% for GPT-4 with plugins at publication — questions conceptually simple for people, hard for tool-using models.",
    "url": "https://agentreliability.dev/k/gaia-benchmark",
    "license": "https://spdx.org/licenses/CC-BY-4.0.html",
    "dateModified": "2026-08-08",
    "citation": [
      {
        "@type": "CreativeWork",
        "name": "GAIA: a benchmark for General AI Assistants",
        "url": "https://arxiv.org/abs/2311.12983"
      }
    ],
    "author": {
      "@type": "Person",
      "name": "Santiago Santa María Morales",
      "jobTitle": "practitioner — harness engineering and agent evaluation in production"
    },
    "@type": "Article",
    "headline": "GAIA benchmark"
  }
}