Text-to-video generation on VBench (Detailed Consistency and Quality Metrics)
98.75Subject ConsistencyPARE + Step Distill
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| PARE + Step Distill#Act.Params=7.7B(45%↓), Steps=42026.05 | 98.75 | 98.33 | 67.92 | 69.86 | 98.74 | 28.4 | 77 | |
| LTX-Video#Act.Params=13B, Steps=402026.05 | 98.42 | 97.76 | 66.8 | 69.97 | 99.07 | 23.93 | 75.99 | |
| VideoCrafter2#Act.Params=1.4B, Steps=50×22026.05 | 98.16 | 98.48 | 66.34 | 68.46 | 98.42 | 17.48 | 74.56 | |
| F3-Pruning#Act.Params=10.5B(25%↓), Steps=50×22026.05 | 97.84 | 96.5 | 57.92 | 57.83 | 98.55 | 10.37 | 69.84 | |
| PARE#Act.Params=6.7B(52%↓), Steps=50×22026.05 | 97.56 | 97.37 | 66.39 | 71.55 | 98.76 | 30.98 | 77.1 | |
| PARE (Width-only)#Act.Params=9.8B(30%↓), Steps=50×22026.05 | 97.36 | 97.5 | 67.81 | 71.39 | 98.79 | 31.9 | 77.46 | |
| CausVid#Act.Params=14B, Steps=42026.05 | 97.34 | 97.17 | 68.54 | 69.14 | 98.36 | 34.17 | 77.45 | |
| AccVideo#Act.Params=14B, Steps=52026.05 | 97.13 | 97.85 | 69.32 | 62.29 | 98.85 | 30.06 | 75.92 | |
| Wan2.1-T2V (Teacher)#Act.Params=14B, Steps=50×22026.05 | 96.99 | 97.49 | 68.76 | 68.2 | 98.59 | 36.2 | 77.7 | |
| rCM#Act.Params=14B, Steps=42026.05 | 96.92 | 97.2 | 69.87 | 69.3 | 98.66 | 32.54 | 77.42 | |
| Wan2.1-T2V#Act.Params=1.3B, Steps=50×22026.05 | 96.91 | 97.16 | 66.26 | 67.49 | 98.52 | 33.89 | 76.7 | |
| PARE (Routing-only)#Act.Params=9.5B(32%↓), Steps=50×22026.05 | 96.55 | 97.21 | 66.3 | 70.02 | 97.76 | 35.24 | 77.18 | |
| DMD2#Act.Params=14B, Steps=42026.05 | 95.38 | 96.16 | 68 | 70.54 | 98.36 | 40.66 | 78.18 | |
| NeoDragon#Act.Params=10.5B(25%↓), Steps=50×22026.05 | 94.33 | 96.28 | 59.5 | 69 | 97.44 | 31.54 | 74.68 | |
| CogVideoX1.5#Act.Params=5B, Steps=50×22026.05 | 92.66 | 94.95 | 56.24 | 62.81 | 96.96 | 46.11 | 74.95 | |
| ICMD#Act.Params=9.80B(30%↓), Steps=50×22026.05 | 71.42 | 87.46 | 29.13 | 37.66 | 97.2 | 15.35 | 56.37 |