Scores are joint: the percentage of test prompts whose video both follows the prompt (SA ≥ 4) and is physically plausible (PC ≥ 4). PA and OI are the physical-activity and object-interaction categories. Last updated November 2026.
Human Leaderboard
| # | Model | Type | All | Hard | PA | OI |
| 1 | Wan2.1-14B | Open | 32.6 | 21.9 | 31.5 | 36.2 |
| 2 | CogVideoX-5B | Open | 25.0 | 0.0 | 24.6 | 26.1 |
| 3 | Cosmos-Diff-7B | Open | 24.1 | 10.9 | 22.6 | 27.4 |
| 4 | Sora† | Closed | 23.3 | 5.3 | 22.2 | 26.7 |
| 5 | Ray2† | Closed | 20.3 | 8.3 | 21.0 | 18.5 |
| 6 | Hunyuan-13B | Open | 17.2 | 6.2 | 17.6 | 15.9 |
| 7 | VideoCrafter-2 | Open | 10.5 | 2.9 | 10.1 | 13.1 |
AutoEval Leaderboard
VideoPhy-2-AutoEval scores we computed on the videos from our paper (per-video scores). Videos were generated from the upsampled captions; semantic adherence is scored against the original caption, matching our human evaluation. AutoEval scores are not directly comparable with human scores.
| # | Model | Type | All | Hard |
| 1 | Sora† | Closed | 21.7 | 5.3 |
| 2 | Hunyuan-13B | Open | 19.6 | 6.2 |
| 3 | CogVideoX-5B | Open | 19.5 | 5.1 |
| 4 | Ray2† | Closed | 17.5 | 5.0 |
| 5 | Wan2.1-14B | Open | 16.9 | 2.2 |
| 6 | VideoCrafter-2 | Open | 16.4 | 3.6 |
| 7 | Cosmos-Diff-7B | Open | 14.7 | 3.4 |
† Sora was evaluated on a 60-prompt subset through its UI (no API); Ray2 on 394 prompts due to API budget.
Results Reported in Other Papers
VideoPhy-2-AutoEval joint scores reported by other papers, not verified by us. Their generation and scoring setups vary (for example, which caption semantic adherence is scored against), so these numbers are not directly comparable with the tables above.
| # | Model | Type | All | Hard | Reported in |
| 1 | PhyEvolve | Agentic | 30.1 | 12.8 | PhyEvolve |
| 2 | CogVideoX-5B | Open | 25.9 | 5.0 | PhyEvolve |
| 3 | Wan2.2-TI2V-5B | Open | 24.9 | 7.2 | PhyEvolve |
| 4 | Co-Director | Agentic | 24.9 | 5.0 | PhyEvolve |
| 5 | Minimax-H3 | Closed | 23.9 | 4.4 | PhyEvolve |
| 6 | Veo 3.1 | Closed | 23.7* | – | Physis-Lang |
| 7 | Wan2.1-T2V-14B | Open | 23.5* | – | Physis-Lang |
| 8 | NEWTON | Agentic | 23.2 | 6.7 | PhyEvolve |
| 9 | CogVideoX1.5-5B | Open | 23.0* | – | Physis-Lang |
| 10 | Cosmos3-Nano | Open | 22.8 | – | VeriPhy |
| 11 | Cosmos-Predict2.5 | Open | 22.7 | 7.8 | PhyEvolve |
| 12 | HunyuanVideo | Open | 22.5 | 6.7 | PhyEvolve |
| 13 | LTX-Video-2.5 | Open | 22.2 | 3.9 | PhyEvolve |
| 14 | LTX-Video-2B | Open | 21.4 | 4.4 | PhyEvolve |
* The source does not state the split or metric; we assume the full-test joint score and are confirming with the authors. – means not reported.
To add your model, run VideoPhy-2-AutoEval on the test set and open a pull request adding your self-reported scores with a link to your paper or code. Questions: cipeng@g.ucla.edu.