reasoning benchmark
GAIA2
GAIA2 evaluates general-purpose AI agents on real-world, multi-step questions that require reasoning, tool use, and information retrieval.
Updated Aug 11, 2026
Models1
Model coverage1
MetricScore
EvidenceB
GAIA2 Ranking
Higher score ranks better on this benchmark.
| 01 | ME | 43.3% | 100.0% | 1 | C |
GAIA2 Highlights
The leading models and scores on this benchmark.
What is GAIA2?
What GAIA2 measures and how its scores work.
GAIA2 evaluates general-purpose AI agents on real-world, multi-step questions that require reasoning, tool use, and information retrieval.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- GAIA2
- Modality
- multimodal
- Primary category
- reasoning
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- gaia2|llm-stats-current
Benchmark scores retain their original unit. Overall score eligibility is shown separately.
FAQ
Common questions about GAIA2.
Which model scores highest on GAIA2?
Muse Glimmer-30B is currently ranked first with 43.3%.
What does GAIA2 measure?
GAIA2 evaluates general-purpose AI agents on real-world, multi-step questions that require reasoning, tool use, and information retrieval.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
1 model results are currently shown.
Does this benchmark affect the overall score?
No. This benchmark is shown for reference but does not contribute to the overall score.