CausalPhys: leaderboard
Metric: Answer accuracy (0-1 scaled to %) over all four domains (Perception, Anticipation, Intervention, Goal-Orientation; 3,062 image- and video-based questions with verified answers over real-world physical scenes), item-weighted; higher is better. Source: arxiv.org. Saturation forecast: Around 2030. 11 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | InternVL3-78B | 59.24 |
| 2 | GPT-4o | 58.88 |
| 3 | Gemini 2.5 Flash | 57.25 |
| 4 | GPT-4o Mini | 56.3 |
| 5 | Claude Sonnet 4 | 54.28 |
| 6 | Qwen 3 VL 32B | 52.68 |
| 7 | Phi-4 Multimodal Instruct | 51.99 |
| 8 | Qwen 2 VL 7B | 50.65 |
| 9 | Mistral Small 3.2 | 46.11 |
Interactive version: theaggregate.ai/benchmark?slug=causalphys · How It Works · Data refreshed daily, snapshot 2026-09-29.