RealDataAgentBench: leaderboard
Data-science agent benchmark evaluating whether LLM agents solve real-data analysis tasks correctly and robustly across correctness, code quality, efficiency, and statistical validity.
Metric: Average DAB Score (%). Source: patibandlavenkatamanideep.github.io. Saturation forecast: Around January 2027. 14 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.8 | 88.89 | #37 |
| 2 | GPT-4.1 | 87.52 | #248 |
| 3 | GPT-4.1 Mini | 87.02 | #355 |
| 4 | Claude Sonnet 4.6 | 85.74 | #90 |
| 5 | GPT-4o | 85.13 | #343 |
| 6 | Claude Opus 4.6 | 84.62 | #63 |
| 7 | Grok 3 Mini | 82.69 | #281 |
| 8 | Claude Haiku 4.5 (20251001) | 80.05 | #261 |
| 9 | GPT-4o Mini | 78.49 | #599 |
| 10 | GPT-5 | 77.95 | #98 |
| 11 | Gemini 2.5 Flash | 66.21 | #245 |
| 12 | GPT-4.1 Nano | 62.37 | #731 |
| 13 | gemma-4-E4B-it | 55.54 | #566 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=realdataagentbench · How It Works · Data refreshed daily, snapshot 2026-10-20.