GAIA
Real-world questions for general AI assistants that need tools and browsing
agents and tool useactive
Definition
- Version
- not in the notes
- Category
- agents and tool use
- Maintainer
- Meta, Hugging Face and AutoGPT authors, USA/France
- Status
- active
- Home page
- https://huggingface.co/spaces/gaia-benchmark/leaderboard
Sources
- GAIA: a benchmark for General AI Assistants arxiv.org https://arxiv.org/abs/2311.12983
- huggingface.co/spaces/gaia-benchmark/leaderboard https://huggingface.co/spaces/gaia-benchmark/leaderboard
- GAIA Leaderboard & Scores , October 2026 benchlm.ai https://benchlm.ai/benchmarks/gaia
Scores
| Model | Score | Who ran it | Setting | View or population | Date | Source |
|---|---|---|---|---|---|---|
 Agents-A1-4B | 95.1 % | vendor-reported | Self-report; only BenchLM row, not marked current | BenchLM | – | benchlm.ai |
 GPT-4 | 15 % | maintainer-run | GPT-4 with plugins, at benchmark release | – | Nov 2023 | arxiv.org |
Where this comes from
Scores from different settings, harnesses or dates are not directly comparable; the row says which. Sorted by score, highest first, as a reading aid, not a ranking.
To cite this page: Project AURION, AI Use-Cases Research & Innovation, "GAIA", https://aurion.semai.id/benchmarks/gaia/, records of October 2026.