Text-to-Video Generation on VBench 1.0 (test)
82.1VBench ScoreOriginal (50 steps)
Evaluation Results
| Method | Links | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Original (50 steps)Backbone=HunyuanVideo, Steps=50, Latency=597.92s, Speedup=1.00x2025.12 | 82.1 | — | — | — | — | — | — | — | — | — | — | — | |
| X-Slim(S)-slowBackbone=HunyuanVideo, Strategy=Step-level, Latency=233.56s, Speedup=2.56x2025.12 | 81.92 | — | — | — | — | — | — | — | — | 28.46 | 0.892 | 0.0893 | |
| TeaCacheBackbone=HunyuanVideo, delta=0.15, Latency=254.43s, Speedup=2.35x2025.12 | 81.9 | — | — | — | — | — | — | — | — | 23.18 | 0.8037 | 0.1827 | |
| X-Slim(C2F)-slowBackbone=HunyuanVideo, Strategy=Coarse-to-Fine, Latency=203.37s, Speedup=2.94x2025.12 | 81.87 | — | — | — | — | — | — | — | — | 27.52 | 0.8794 | 0.1062 | |
| X-Slim(S)-fastBackbone=HunyuanVideo, Strategy=Step-level, Latency=183.41s, Speedup=3.26x2025.12 | 81.75 | — | — | — | — | — | — | — | — | 24.61 | 0.8346 | 0.1619 | |
| Uniform CacheBackbone=HunyuanVideo, N=2, Latency=290.25s, Speedup=2.06x2025.12 | 81.73 | — | — | — | — | — | — | — | — | 19.82 | 0.7294 | 0.2708 | |
| X-Slim(C2F)-fastBackbone=HunyuanVideo, Strategy=Coarse-to-Fine, Latency=169.86s, Speedup=3.52x2025.12 | 81.69 | — | — | — | — | — | — | — | — | 24.07 | 0.8216 | 0.1638 | |
| 50%: 25 stepsBackbone=HunyuanVideo, Steps=25, Latency=301.98s, Speedup=1.98x2025.12 | 81.68 | — | — | — | — | — | — | — | — | 19.94 | 0.7367 | 0.2673 | |
| Uniform CacheBackbone=HunyuanVideo, N=3, Latency=196.04s, Speedup=3.05x2025.12 | 81.52 | — | — | — | — | — | — | — | — | 17.61 | 0.6717 | 0.3524 | |
| EasyCacheBackbone=HunyuanVideo, delta=0.035, Latency=231.75s, Speedup=2.58x2025.12 | 81.51 | — | — | — | — | — | — | — | — | 30.56 | 0.8977 | 0.0814 | |
| 30%: 15 stepsBackbone=HunyuanVideo, Steps=15, Latency=179.56s, Speedup=3.33x2025.12 | 81.3 | — | — | — | — | — | — | — | — | 17.33 | 0.6674 | 0.3692 | |
| TeaCacheBackbone=HunyuanVideo, delta=0.25, Latency=179.12s, Speedup=3.34x2025.12 | 81.3 | — | — | — | — | — | — | — | — | 23.1 | 0.7946 | 0.1973 | |
| EasyCacheBackbone=HunyuanVideo, delta=0.055, Latency=188.62s, Speedup=3.17x2025.12 | 80.77 | — | — | — | — | — | — | — | — | 27.39 | 0.8357 | 0.1745 | |
| AnimateDiffBase Model=AnimateDiff2024.11 | — | 0.9488 | 0.9755 | 0.9228 | 0.9578 | 0.47 | 0.2532 | — | — | — | — | — | |
| CogVideoXModel=CogVideoX2025.12 | — | — | — | — | — | — | — | 0.655 | 0.468 | — | — | — | |
| HunyuanVideoModel=HunyuanVideo2025.12 | — | — | — | — | — | — | — | 0.736 | 0.663 | — | — | — | |
| LavieBase Model=Lavie2024.11 | — | 0.9599 | 0.9739 | 0.9487 | 0.969 | 0.515 | 0.2506 | — | — | — | — | — | |
| ModelScopeModel=ModelScope2025.12 | — | — | — | — | — | — | — | 0.726 | 0.34 | — | — | — | |
| MotionPromptBase Model=Lavie2024.11 | — | 0.9646 | 0.9781 | 0.9625 | 0.9765 | 0.3863 | 0.2415 | — | — | — | — | — | |
| MotionPromptBase Model=AnimateDiff2024.11 | — | 0.9528 | 0.9763 | 0.9258 | 0.9599 | 0.4125 | 0.2529 | — | — | — | — | — | |
| MotionPromptBase Model=VideoCrafter22024.11 | — | 0.9745 | 0.9774 | 0.9588 | 0.9759 | 0.3938 | 0.2451 | — | — | — | — | — | |
| Open-SoraModel=Open-Sora2025.12 | — | — | — | — | — | — | — | 0.684 | 0.477 | — | — | — | |
| VideoCrafter2Base Model=VideoCrafter22024.11 | — | 0.9736 | 0.9559 | 0.9559 | 0.975 | 0.4088 | 0.2498 | — | — | — | — | — |