Text-to-Video Generation on VBench 2.0
64.16Overall ScoreWan2.1
Evaluation Results
| Method | Links | ||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Wan2.1Denoising Steps=20, Model=Wan2.1-1.3B2026.03 | 64.16 | — | — | — | — | — | — | — | — | — | — | 88 | 3,568.83 | — | — | — | |
| TeacacheCache Type=Reuse-based, Speed Setting=slow2026.03 | 60.73 | — | — | — | — | — | — | — | — | — | — | 75 | 3,076.39 | 27.19 | 86.74 | 10.73 | |
| TC-PadéCache Type=Prediction-based, Speed Setting=fast2026.03 | 60.38 | — | — | — | — | — | — | — | — | — | — | 51 | 2,055.24 | 21.7 | 63.87 | 30.01 | |
| TeacacheCache Type=Reuse-based, Speed Setting=fast2026.03 | 58.4 | — | — | — | — | — | — | — | — | — | — | 61 | 2,583.95 | 21.35 | 61.42 | 32.65 | |
| Diffusion-DRFReward Model=Qwen2.5-VL-7B2026.01 | 55.38 | 54.58 | 56.96 | 27.98 | 80.51 | 56.85 | 75.82 | 42.86 | 40.23 | 21.56 | 24.69 | — | — | — | — | — | |
| TaylorSeerCache Type=Prediction-based, Hyperparameters=N = 3, O = 12026.03 | 54.74 | — | — | — | — | — | — | — | — | — | — | 67 | 1,954.23 | 15.02 | 35.96 | 56.57 | |
| TaylorSeerCache Type=Prediction-based, Hyperparameters=N = 4, O = 12026.03 | 54.5 | — | — | — | — | — | — | — | — | — | — | 53 | 1,876.24 | 14.93 | 35.32 | 58.56 | |
| Diffusion-DRF-miniReward Model=Qwen2.5-VL-3B2026.01 | 53.72 | 53.93 | 57.53 | 25.95 | 76.74 | 54.42 | 70 | 42.49 | 41.38 | 18.22 | 23.15 | — | — | — | — | — | |
| Pre-trainedBackbone=Wan2.1-3B-T2V2026.01 | 52.99 | 53.79 | 55.52 | 26.59 | 80.65 | 48.4 | 36.23 | 37 | 37.36 | 17.33 | 20.68 | — | — | — | — | — | |
| VideoAlign*Reward Model=VideoAlign, Framework=Fine-tuned under the same setting2026.01 | 52.84 | 49.87 | 55.81 | 27.41 | 78.47 | 52.66 | 65.26 | 41.76 | 37.93 | 18.44 | 23.15 | — | — | — | — | — | |
| Phantom2026.04 | 51.84 | 45.51 | 61.43 | 20.23 | 88.39 | 43.61 | — | — | — | — | — | — | — | — | — | — | |
| Wan2.2-TI2V-5BModel size=5B2026.04 | 51.57 | 52.5 | 60.57 | 18.5 | 86.1 | 40.19 | — | — | — | — | — | — | — | — | — | — | |
| Flow-GRPOBackbone=Wan2.1-3B-T2V2026.01 | 50.64 | 44.71 | 50.85 | 25.48 | 77.8 | 54.37 | 69.07 | 36.63 | 36.21 | 14.89 | 19.32 | — | — | — | — | — | |
| PickScore*Reward Model=PickScore, Framework=Fine-tuned under the same setting2026.01 | 49.62 | 35.97 | 55.23 | 23.88 | 81.89 | 51.13 | 66.67 | 39.19 | 36.78 | 16.89 | 22.84 | — | — | — | — | — |