BehaviorBench - Scientific Workflow Prediction: leaderboard
Metric: BLEURT score of generated research-workflow aspects (key idea, method, outcome, projected impact and title, each from the preceding aspects) against 2025 American Economic Review and Nature Human Behaviour papers; higher is better. Source: arxiv.org. Saturation forecast: Around August 2027. 18 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Opus 4.6 | 0.48 |
| 2 | Claude Sonnet 4.6 | 0.47 |
| 3 | Gemini 3.1 Pro (Preview) | 0.47 |
| 4 | GPT-5.4 (High) | 0.46 |
| 5 | GPT-4.1 | 0.46 |
| 6 | GPT-5.4 Mini (High) | 0.45 |
| 7 | Qwen 3 4B | 0.45 |
| 8 | Gemini 3.1 Flash Lite (Preview) | 0.43 |
| 9 | Llama 3.3 70B Instruct | 0.43 |
| 10 | DeepSeek V3.2 | 0.43 |
| 11 | Claude Haiku 4.5 | 0.43 |
Interactive version: theaggregate.ai/benchmark?slug=behaviorbench-scientific-workflow-prediction · How It Works · Data refreshed daily, snapshot 2026-09-29.