CausalPhys - Anticipation: leaderboard
Metric: Answer accuracy (0-1 scaled to %) on the Anticipation domain (predicting physical outcomes such as collisions, deformation and fluid flow) of the image- and video-based questions with verified answers over real-world physical scenes; higher is better. Source: arxiv.org. Saturation forecast: Around 2031. 11 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-4o | 60.11 |
| 2 | GPT-4o Mini | 59.11 |
| 3 | Gemini 2.5 Flash | 58.22 |
| 4 | InternVL3-78B | 58 |
| 5 | Phi-4 Multimodal Instruct | 55.33 |
| 6 | Claude Sonnet 4 | 53.22 |
| 7 | Qwen 2 VL 7B | 52.22 |
| 8 | Qwen 3 VL 32B | 51.89 |
| 9 | Mistral Small 3.2 | 41 |
Interactive version: theaggregate.ai/benchmark?slug=causalphys-anticipation · How It Works · Data refreshed daily, snapshot 2026-09-29.