BilliardPhys-Bench: leaderboard
Current multimodal models handle static image recognition well, but intuitive physical reasoning remains a weakness. Predicting how objects will move and interact from a single image is still difficult for these systems. We present BilliardPhys-Bench, a benchmark for physical reasoning in synthetic billiards environments. Its procedural engine generates randomized scenarios with friction and elast
Metric: Total (Weighted) (self-reported). Source: benchmarklist.com. Saturation forecast: Around August 2027. 13 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.5 | 73.58 | #31 |
| 2 | GPT-5.4 Pro (xHigh) | 72.29 | #32 (GPT-5.4 Pro) |
| 3 | GPT-5.2 Pro | 62.34 | #61 |
| 4 | Gemini 3 Flash (Preview) | 58.06 | #82 |
| 5 | Qwen 3.6 Plus | 57.36 | #120 |
| 6 | Gemini 3.1 Pro (Preview) | 53.84 | #58 |
| 7 | Seed 2.0 Pro | 48.46 | #118 |
| 8 | Claude Opus 4.6 | 44.42 | #63 |
| 9 | Qwen 3.5 Plus | 43.09 | #125 |
| 10 | Kimi K2.5 | 41.95 | #144 |
| 11 | Claude Opus 4.7 | 39.38 | #49 |
| 12 | Seed 1.8 | 37.26 | #139 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=billiardphys-bench · How It Works · Data refreshed daily, snapshot 2026-10-20.