Text-to-Video Generation on VBench (Performance and Quality Metrics)
82.9Latency (s)Spectrum
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| SpectrumBase Model=HunyuanVideo, α=3.02026.03 | 82.9 | 4.56 | 25.39 | 77.9 | 0.273 | 83.89 | |
| FORABase Model=HunyuanVideo, N=62026.03 | 95.77 | 3.95 | 17.29 | 59.4 | 0.475 | 83.35 | |
| TaylorSeerBase Model=HunyuanVideo, N=6, O=12026.03 | 97.83 | 3.86 | 22.27 | 74 | 0.303 | 83.4 | |
| TeaCacheBase Model=HunyuanVideo, δ=0.252026.03 | 102.52 | 3.69 | 21.42 | 72.5 | 0.334 | 83.68 | |
| SpectrumBase Model=Wan2.1-14B, α=3.02026.03 | 104.18 | 4.67 | 21.24 | 69.4 | 0.265 | 82.21 | |
| SpectrumBase Model=HunyuanVideo, α=0.752026.03 | 112.59 | 3.36 | 27.77 | 84.2 | 0.209 | 84.11 | |
| 15 stepsBase Model=HunyuanVideo, Steps=152026.03 | 119.58 | 3.16 | 17.4 | 62 | 0.459 | 83.21 | |
| FORABase Model=Wan2.1-14B, N=62026.03 | 120.92 | 4.02 | 13.02 | 35.3 | 0.598 | 80.82 | |
| FORABase Model=HunyuanVideo, N=42026.03 | 122.42 | 3.09 | 18.24 | 66.3 | 0.427 | 83.34 | |
| TaylorSeerBase Model=Wan2.1-14B, N=6, O=12026.03 | 123.49 | 3.94 | 17.24 | 58.5 | 0.367 | 81.38 | |
| TaylorSeerBase Model=HunyuanVideo, N=4, O=12026.03 | 127.55 | 2.96 | 24.73 | 80.5 | 0.243 | 84.03 | |
| TeaCacheBase Model=Wan2.1-14B, δ=0.252026.03 | 127.59 | 3.81 | 16.53 | 55.7 | 0.39 | 81.85 | |
| TeaCacheBase Model=HunyuanVideo, δ=0.22026.03 | 130.44 | 2.9 | 23.88 | 78.3 | 0.27 | 84.08 | |
| SpectrumBase Model=Wan2.1-14B, α=0.752026.03 | 143.03 | 3.4 | 22.78 | 74.9 | 0.222 | 82.8 | |
| 15 stepsBase Model=Wan2.1-14B, Steps=152026.03 | 150.5 | 3.23 | 13.92 | 39.7 | 0.546 | 80.77 | |
| FORABase Model=Wan2.1-14B, N=42026.03 | 155.8 | 3.12 | 14.47 | 42.6 | 0.532 | 80.53 | |
| TaylorSeerBase Model=Wan2.1-14B, N=4, O=12026.03 | 161.5 | 3.01 | 19.46 | 66 | 0.299 | 82.74 | |
| TeaCacheBase Model=Wan2.1-14B, δ=0.22026.03 | 164.78 | 2.95 | 19.13 | 62.8 | 0.347 | 82.59 | |
| 25 stepsBase Model=HunyuanVideo, Steps=252026.03 | 193.19 | 1.96 | 19.64 | 67.3 | 0.376 | 84.78 | |
| ToCaBase Model=HunyuanVideo, N=6, R=0.92026.03 | 199.1 | 1.9 | 19.48 | 65.6 | 0.438 | 83.62 | |
| 25 stepsBase Model=Wan2.1-14B, Steps=252026.03 | 246.3 | 1.97 | 14.94 | 45.9 | 0.481 | 81.74 | |
| ToCaBase Model=HunyuanVideo, N=4, R=0.92026.03 | 247.25 | 1.53 | 21.26 | 73.2 | 0.357 | 83.79 | |
| ToCaBase Model=Wan2.1-14B, N=6, R=0.92026.03 | 250.57 | 1.94 | 15.1 | 48.8 | 0.453 | 81.14 | |
| ToCaBase Model=Wan2.1-14B, N=4, R=0.92026.03 | 313.62 | 1.55 | 16.88 | 53.7 | 0.41 | 82.18 | |
| 50 stepsBase Model=HunyuanVideo, Steps=502026.03 | 378.3 | 1 | — | — | — | 84.61 | |
| 50 stepsBase Model=Wan2.1-14B, Steps=502026.03 | 486.12 | 1 | — | — | — | 83.15 |