Controllable Video Generation on VBench (test)
99.4Motion SmoothnessLess is More
Evaluation results
| Method | Links | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Less is MoreControl Type=Aperture (AP), Inference Strategy=Joint (dirty)2025.11 | 99.4 | 25.181 | 0.427 | 96.8 | 97.9 | 21.9 | 62.9 | 59.6 | — | — | |
| AnimateAnyMeshText condition=true, Trajectory condition=false2026.06 | 99.2 | — | — | — | — | 58.5 | 53.9 | — | — | 16.8 | |
| BiMotionText condition=true, Trajectory condition=false2026.06 | 99.1 | — | — | — | — | 32.9 | 53.6 | — | — | 17 | |
| Less is MoreControl Type=Temperature (TP), Inference Strategy=Joint (dirty)2025.11 | 99 | 25.487 | 0.55 | 96 | 97.5 | 40.6 | 62 | 66.4 | — | — | |
| Less is MoreControl Type=Temperature (TP), Inference Strategy=Decoupled (clean)2025.11 | 98.9 | 25.595 | 0.532 | 95 | 97.2 | 41.7 | 60.5 | 62.3 | — | — | |
| Baseline (WAN 2.1)Mode=Baseline2025.11 | 98.8 | 25.39 | 0.522 | 95.1 | 97.3 | 42.7 | 60.6 | 61.8 | — | — | |
| speed-conditioned video generation2026.04 | 98.71 | — | — | — | — | — | 53.94 | 63.37 | 97 | — | |
| Less is MoreControl Type=Shutter (ST), Inference Strategy=Decoupled (clean)2025.11 | 98.7 | 25.587 | 0.521 | 94.6 | 96.9 | 46.9 | 60.4 | 59.6 | — | — | |
| Less is MoreControl Type=Aperture (AP), Inference Strategy=Decoupled (clean)2025.11 | 98.7 | 25.595 | 0.513 | 95.1 | 97.3 | 43.8 | 61 | 63.3 | — | — | |
| OursText condition=true, Trajectory condition=true2026.06 | 98.4 | — | — | — | — | 74.3 | 54.4 | — | — | 17.5 | |
| Less is MoreControl Type=Shutter (ST), Inference Strategy=Joint (dirty)2025.11 | 98.3 | 25.295 | 0.453 | 93.9 | 96.3 | 68.8 | 57.6 | 53.1 | — | — | |
| Wan2.12026.04 | 98.03 | — | — | — | — | — | 53.84 | 63.67 | 96.38 | — |