Physical Reasoning from Video — leaderboard
Video physical-reasoning leaderboard reporting model scores on IntPhys 2, MVPBench, and CausalVQA-style physical world understanding tasks.
Metric: Average Accuracy (%). Source: huggingface.co. Status: years away from saturation. 8 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 2.5 Flash | 58.88 |
| 2 | GPT-4o | 45.55 |
| 3 | Gemini 1.5 Pro | 40.85 |
Interactive version: theaggregate.ai/benchmark?slug=physical-reasoning-from-video · How the rankings work · Data refreshed daily, snapshot 2026-07-22.