{
  "generated": "2026-09-08",
  "benchmark": "gaia",
  "results": [
    {
      "system": "GPT4 + plugins",
      "developer": "OpenAI",
      "value": 15.0,
      "date": "2023-11-21",
      "source": {
        "url": "https://arxiv.org/abs/2311.12983",
        "kind": "paper",
        "accessed": "2026-09-04"
      },
      "conditions": {
        "tools": true,
        "notes": "Headline figure from the abstract; Table 4 gives 30.3 / 9.7 / 0 by level."
      }
    },
    {
      "system": "HAL Generalist Agent + Claude Sonnet 4.5",
      "developer": "Anthropic",
      "value": 74.55,
      "date": "2025-09-01",
      "source": {
        "url": "https://hal.cs.princeton.edu/gaia",
        "kind": "independent-evaluation",
        "accessed": "2026-09-04"
      },
      "conditions": {
        "split": "validation",
        "scaffold": "HAL Generalist Agent",
        "tools": true,
        "notes": "HAL re-run on the validation set, 2 runs, $178 total cost. Month-only date on the leaderboard; 1st used."
      }
    },
    {
      "system": "Nemotron-ToolOrchestra",
      "developer": "NVIDIA",
      "value": 87.38,
      "date": "2025-12-02",
      "source": {
        "url": "https://huggingface.co/spaces/gaia-benchmark/leaderboard",
        "kind": "official-leaderboard",
        "accessed": "2026-09-04"
      },
      "conditions": {
        "split": "test",
        "scaffold": "ToolOrchestra (Nemotron-ToolOrchestrator-8B orchestrating GPT-5, Claude Opus 4.1, Qwen2.5-Math-72B)",
        "tools": true
      }
    },
    {
      "system": "Co-Sight Pro v1.0.1",
      "developer": "ZTE-AICloud",
      "value": 93.02,
      "date": "2026-05-16",
      "source": {
        "url": "https://huggingface.co/spaces/gaia-benchmark/leaderboard",
        "kind": "official-leaderboard",
        "accessed": "2026-09-04"
      },
      "conditions": {
        "split": "test",
        "scaffold": "Co-Sight Pro",
        "tools": true,
        "notes": "Ensemble of ZTE Nebula LLM, Gemini 3.1 Pro, GPT 5.5 and Claude Opus 4.7 (self-submitted, auto-scored)."
      }
    },
    {
      "system": "CustomGPT.ai Research Lab v44",
      "developer": "CustomGPT.ai",
      "value": 93.36,
      "date": "2026-06-03",
      "source": {
        "url": "https://huggingface.co/spaces/gaia-benchmark/leaderboard",
        "kind": "official-leaderboard",
        "accessed": "2026-09-04"
      },
      "conditions": {
        "split": "test",
        "scaffold": "CustomGPT.ai enterprise agent",
        "tools": true,
        "notes": "Top of the test leaderboard as of access date; ensemble of Claude, Gemini and GPT models (self-submitted, auto-scored)."
      }
    }
  ]
}