Text-to-Video Generation on MovieGen-Bench 60s 100 text prompts long-form evaluation
91.45Dino Score (10s)Ours, student
Evaluation results
| Method | Links | ||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Ours, studentContext Length=20+s2026.02 | 91.45 | 89.25 | 89.66 | 87.45 | 88.33 | 87.89 | 95.82 | 94.75 | 95.05 | 93.88 | 94.92 | 95.35 | 37.12 | 36.25 | 36.72 | 37.15 | 37.08 | 37.66 | 95.95 | 95.68 | |
| LongLiveContext Length=3.0s2026.02 | 91.25 | 89.55 | 89.12 | 86.51 | 87.83 | 86.26 | 95.74 | 94.25 | 92.8 | 91.5 | 93.12 | 94.82 | 36.95 | 35.8 | 36.17 | 36.58 | 35.92 | 37.13 | 94.92 | 93.05 | |
| Infinate RoPEContext Length=1.5s2026.02 | 91.18 | 88.17 | 85.37 | 79.8 | 81.1 | 83.72 | 94.09 | 91.71 | 91.13 | 89.71 | 86.11 | 88.88 | 35.26 | 35.03 | 35.88 | 32.56 | 32.29 | 32.28 | 92.42 | 90.11 | |
| Ours, teacherContext Length=20+s2026.02 | 87.61 | — | — | — | — | — | 95.52 | — | — | — | — | — | 35.93 | — | — | — | — | — | 95.24 | 94.87 | |
| FramePack-F1Context Length=9.2s2026.02 | 81.86 | 77.95 | 78.84 | 73.1 | 69.52 | 68.5 | 95.41 | 93.42 | 93.92 | 92.44 | 92.47 | 89.36 | 36.36 | 34.67 | 33.75 | 33.84 | 34.77 | 32.3 | 91.61 | 89.15 |