Text-to-Video Generation on VBench Wan 480p
85.05Total ScoreWan2.1 T2V + rCM
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Wan2.1 T2V + rCM#Params=14B, Resolution=832 × 480 × 81, NFE=2, Throughput (FPS)=8.32025.10 | 85.05 | 85.57 | 82.95 | |
| Wan2.1 T2V + rCM#Params=14B, Resolution=832 × 480 × 81, NFE=4, Throughput (FPS)=4.52025.10 | 84.92 | 85.43 | 82.88 | |
| Wan2.1 T2V + DMD2#Params=1.3B, Resolution=832 × 480 × 81, NFE=4, Throughput (FPS)=14.62025.10 | 84.56 | 85.58 | 80.5 | |
| Wan2.1 T2V + rCM#Params=1.3B, Resolution=832 × 480 × 81, NFE=4, Throughput (FPS)=14.62025.10 | 84.43 | 85.38 | 80.63 | |
| Wan2.1 T2V + rCM#Params=1.3B, Resolution=832 × 480 × 81, NFE=2, Throughput (FPS)=232025.10 | 84.09 | 84.9 | 80.86 | |
| Wan2.1 T2V#Params=14B, Resolution=832 × 480 × 81, NFE=50 × 2, Throughput (FPS)=0.182025.10 | 83.58 | 84.26 | 80.92 | |
| Wan2.1 T2V#Params=1.3B, Resolution=832 × 480 × 81, NFE=50 × 2, Throughput (FPS)=0.72, Retested with Diffusers and augmented prompts=true2025.10 | 83.02 | 83.95 | 79.26 | |
| Wan2.1 T2V + rCM#Params=14B, Resolution=832 × 480 × 81, NFE=1, Throughput (FPS)=14.42025.10 | 83.02 | 83.57 | 80.81 | |
| Wan2.1 T2V + rCM#Params=1.3B, Resolution=832 × 480 × 81, NFE=1, Throughput (FPS)=32.32025.10 | 82.65 | 83.6 | 78.82 | |
| Wan2.1Training Paradigm=Multi-step DiT, NFE=40×2, Latency (s)=732, Speedup=1.0×, #Params (B)=14B2026.03 | 82.34 | 83.77 | 76.65 | |
| Wan2.1Training Paradigm=Multi-step DiT, NFE=40×2, Latency (s)=147, Speedup=4.97×, #Params (B)=1.3B2026.03 | 82.11 | 83.77 | 75.46 | |
| SANA-VideoTraining Paradigm=Multi-step DiT, NFE=40×2, Latency (s)=79.03, Speedup=9.26×, #Params (B)=2B2026.03 | 81.93 | 84.11 | 73.23 | |
| Wan2.1 + VSA⋆Training Paradigm=Few-step DiT with teacher-distillation, NFE=4, Latency (s)=11.28, Speedup=64.9×, #Params (B)=1.3B2026.03 | 81.28 | 83.36 | 72.97 | |
| EFlowTraining Paradigm=Few-step Efficient DiT from-scratch, NFE=4, Latency (s)=10.09, Speedup=72.5×, #Params (B)=1.5B2026.03 | 79.69 | 83.53 | 64.34 | |
| Wan2.1 + rCM†Training Paradigm=Few-step DiT with teacher-distillation, NFE=4, Latency (s)=12.61, Speedup=58.0×, #Params (B)=1.3B2026.03 | 79.36 | 83.75 | 61.77 |