Video Generation on EVD-Bench
89.7Human Eval: Text FidelityDiT-30B + EVD
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| DiT-30B + EVDnumber of parameters=30B, module=EVD2026.03 | 89.7 | 92.4 | 97.1 | 78.1 | 95.7 | |
| DiT-4B + EVDParameters=4B, EVD integration=true2026.03 | 88.9 | 91.3 | 96.4 | 76.2 | 94.8 | |
| CogVideo2BParameters=2B2026.03 | 80.2 | 88.1 | 89.7 | 69.8 | 87.6 | |
| PyramidFlow2026.03 | 75.8 | 82.4 | 81.1 | 73.9 | 88.5 | |
| DiT-30Bnumber of parameters=30B2026.03 | 72.4 | 76.1 | 79.5 | 73.8 | 88.7 | |
| DiT-4BParameters=4B2026.03 | 70.6 | 76.8 | 80.3 | 75.4 | 78.9 | |
| Veo 3.12026.03 | 66.9 | 73.5 | 78.4 | 77.2 | 91.8 | |
| CogVideo5BParameters=5B2026.03 | 65.4 | 73.8 | 72.5 | 72.3 | 89.2 | |
| Runway Gen-4.52026.03 | 64.7 | 71.2 | 76.8 | 76.9 | 91.4 | |
| Sora 2 Pro2026.03 | 63.5 | 69.8 | 74.1 | 77.8 | 90.9 | |
| Kling 3.0 Pro2026.03 | 61.8 | 67.4 | 72.9 | 78.6 | 92.6 | |
| Mochi 12026.03 | 58.2 | 63.7 | 70.3 | 72.6 | 88.8 |