Benchmark
Gaia2 A2A
Agentic benchmarks measure how well models carry out tasks that take several steps. Scores are percentages.
- Models measured
- 12
- Scores recorded
- 16
- Sources
- 1
- IQ weight
- 1.51
Score distribution
- 15.9%
- Mean
- 16%
- Median
- 7.6%
- Std dev
- 9.2%
- 10th percentile
- 27.9%
- 90th percentile
- 4.6%
- Min
This benchmark has a weight of 1.51 for chat models, measured over 12 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.
Leaderboard
| # | Model | Score |
|---|---|---|
| 1 | Claude Sonnet 4thinking· Gaia2 (Meta ARE) data | 32.5% |
| 2 | Claude Sonnet 4· Gaia2 (Meta ARE) data | 27.9% |
| 3 | GPT 5low· Gaia2 (Meta ARE) data | 24.6% |
| 4 | Gemini 2.5 Pro· Gaia2 (Meta ARE) data | 20.4% |
| 5 | DeepSeek V3.1 Terminus· Gaia2 (Meta ARE) data | 19.4% |
| 6 | Kimi K2· Gaia2 (Meta ARE) data | 18.3% |
| 7 | GPT 5high· Gaia2 (Meta ARE) data | 17.9% |
| 8 | DeepSeek V3.1· Gaia2 (Meta ARE) data | 16% |
| 9 | Grok 4· Gaia2 (Meta ARE) data | 14.4% |
| 10 | Qwen3 235B (Web)thinking· Gaia2 (Meta ARE) data | 12.5% |
| 11 | GPT 5minimal· Gaia2 (Meta ARE) data | 11.5% |
| 12 | GPT OSS 120Bhigh· Gaia2 (Meta ARE) data | 10.6% |
| 13 | Qwen3 235B (Web)· Gaia2 (Meta ARE) data | 9.4% |
| 14 | Llama 4 Maverick· Gaia2 (Meta ARE) data | 9.2% |
| 15 | GPT 4o· Gaia2 (Meta ARE) data | 5.2% |
| 16 | Llama 3.3 70B Instruct· Gaia2 (Meta ARE) data | 4.6% |