VideoPhy
Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Video Generation | VideoPhy | SA (%)72 | 50 | |
| Text-to-Video Generation | VideoPhy | PC Score40.1 | 41 | |
| AI-Generated Video Detection | VideoPhy 1.0 (test) | CVX Score93.37 | 28 | |
| Physical Plausibility Evaluation | VideoPhy | Average PC41 | 16 | |
| Text-to-Video Generation | VideoPhy2 (ALL) | PC Score72.54 | 16 | |
| Video Detection | VideoPhy | Accuracy89.61 | 14 | |
| AI-Generated Video Detection | VideoPhy | CVX AUC93.37 | 14 | |
| Video Generation | VideoPhy Fluid-Fluid | SA and PC Score55.4 | 11 | |
| Video Generation | VideoPhy Solid-Fluid | SA and PC Score60 | 11 | |
| Video Generation | VideoPhy Solid-Solid | SA and PC Score40.6 | 11 | |
| Video Generation | VideoPhy Overall | SA and PC Score49.3 | 11 | |
| Video Generation | VideoPhy2 (test) | SA46.2 | 11 | |
| Text-to-Video Prompt Rewriting | VideoPhy 2 | SA Score48.4 | 11 | |
| Prompt Enhancement for Video Generation | VideoPhy2 held-out (val) | SA (%)41.8 | 11 | |
| Text-to-Video Generation | VideoPhy2 HARD | PC Score52.8 | 10 | |
| Text-to-Video Generation | VideoPhy-2 | SA Score28.86 | 9 | |
| Video Physical Commonsense Evaluation | VideoPhy-2 | Spearman PC0.76 | 9 | |
| Physical Reasoning | VideoPhy 2 | Accuracy0.386 | 8 | |
| Text-to-Video Generation | VideoPhy2 (test) | Hard Score0.05 | 8 | |
| Physical Realism Evaluation | VideoPhy-2 (test) | SA Score76 | 7 | |
| Video Generation Evaluation | VideoPhy2 | SA Score3.82 | 4 | |
| Human Preference Evaluation | VideoPhy 1.0 (test) | Physics Plausibility Win Rate59.3 | 4 | |
| Video Generation | VideoPhy2 hard and easy | PC (Gemini3-F)55.6 | 4 | |
| Video Generation | VideoPhy2 | SA Score0.29 | 3 |
Showing 24 of 24 rows