Physical Reasoning - IntPhys 2: leaderboard
Physical-reasoning video slice measuring detection of physical consistency and plausibility in IntPhys 2 examples.
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 5 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.5 Flash | 56.1 | #245 |
| 2 | GPT-4o | 53.19 | #343 |
| 3 | Gemini 1.5 Pro | 52.1 | #445 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=physical-reasoning-intphys-2 · How It Works · Data refreshed daily, snapshot 2026-10-20.