| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| VBench | Phantom-1.3B | Motion Smoothness98.93 | 8 | 2mo ago | |
| Subject-driven Video Generation | CustomCrafter | Training Steps10 | 7 | 2mo ago | |
| 3D-CustomBench | VACE | CLIP-I0.8964 | 5 | 4mo ago | |
| Subject-to-Video (S2V) (test) | LibraGen | MS0.538 | 5 | 4mo ago | |
| Human-Object Interaction (HOI) scenes | MV-S2V | DINO (v -> r)71.7 | 5 | 5mo ago | |
| Object-Centric (OC) scenes | MV-S2V | DINO Sim (V->R)0.816 | 5 | 5mo ago | |
| S2V 4 reference images (test) | LibraGen | GSB Ratio70 | 4 | 4mo ago | |
| S2V 3 reference images (test) | LibraGen | GSB Ratio0.429 | 4 | 4mo ago | |
| S2V 2 reference images (evaluation set) | LibraGen | GSB Ratio0.62 | 4 | 4mo ago | |
| S2V 1 reference image (val) | LibraGen | GSB Ratio42.3 | 4 | 4mo ago |
