Robotic Video Generation on R-Bench
63.8Average ScorePF-Cosmos
Evaluation results
| Method | Links | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| PF-CosmosModel Category=Finetuned methods, Backbone=Cosmos3-Nano, Fine-tuning status=ft2026.06 | 63.8 | 58.9 | 69.7 | 51.3 | 60.8 | 53 | 69.3 | 70 | 72.2 | 69 | |
| PF-WanModel Category=Finetuned methods, Backbone=Wan2.2-I2V-A14B, Fine-tuning status=ft2026.06 | 62 | 56.4 | 65.4 | 49.1 | 58.4 | 52.4 | 68.7 | 69.6 | 69.2 | 68.5 | |
| LingBot-VideoType=open-source2026.07 | 62 | 57.8 | 64.3 | 44.4 | 63.4 | 50.5 | 63.6 | 63.9 | 75.8 | 68.9 | |
| Cosmos3-NanoModel Category=Finetuned methods, Fine-tuning status=ft2026.06 | 61.5 | 57.8 | 67.4 | 49.2 | 59.3 | 48.5 | 66.5 | 67.1 | 70.6 | 67.1 | |
| Wan2.6Model Category=Commercial2026.06 | 60.7 | 54.6 | 65.6 | 47.9 | 51.4 | 53.1 | 66.6 | 68.1 | 72.3 | 66.7 | |
| Wan 2.6Type=closed-source2026.07 | 60.7 | 54.6 | 65.6 | 47.9 | 51.4 | 53.1 | 66.6 | 68.1 | 72.3 | 66.7 | |
| Veo 3.1Model Category=Commercial2026.06 | 59.9 | 54.1 | 47.4 | 53.4 | 59.2 | 46.7 | 67 | 66.6 | 74.3 | 70.4 | |
| Seedance 1.5 ProModel Category=Commercial2026.06 | 58.4 | 57.7 | 49.5 | 48.4 | 57 | 47 | 64.8 | 64.1 | 68 | 69.2 | |
| Cosmos3-NanoModel Category=Finetuned methods, Fine-tuning status=Original2026.06 | 58.4 | 55 | 67 | 46.6 | 57.6 | 39.4 | 59.1 | 61.1 | 73.1 | 66.6 | |
| Seedance 1.5 proType=closed-source2026.07 | 58.4 | 57.7 | 49.5 | 48.4 | 57 | 47 | 64.8 | 64.1 | 68 | 69.2 | |
| Cosmos3-SuperModel Category=Robotics-specific2026.06 | 58.1 | 48.7 | 64.2 | 44.4 | 59.1 | 39.5 | 61.5 | 62.3 | 73.9 | 69.1 | |
| Cosmos3 SuperType=open-source2026.07 | 58.1 | 48.7 | 64.2 | 44.4 | 59.1 | 39.5 | 61.5 | 62.3 | 73.9 | 69.1 | |
| Wan2.2-I2V-A14BModel Category=Finetuned methods, Fine-tuning status=ft2026.06 | 57.9 | 52.3 | 62.8 | 45.2 | 54.5 | 47.8 | 64.2 | 63.5 | 65.6 | 65.3 | |
| Wan2.5Model Category=Commercial2026.06 | 57 | 52.7 | 57.6 | 40.2 | 49.6 | 43.7 | 68 | 63.4 | 72.6 | 65.4 | |
| Hailuo v2Model Category=Commercial2026.06 | 56.5 | 56 | 63.7 | 38.6 | 54.5 | 47.4 | 59.4 | 61.1 | 64 | 63.5 | |
| Veo 3Model Category=Commercial2026.06 | 56.3 | 52.1 | 50.8 | 43 | 53 | 50.4 | 63.4 | 61 | 68.9 | 63.7 | |
| Veo 3Type=closed-source2026.07 | 56.3 | 52.1 | 50.8 | 43 | 53 | 50.4 | 63.4 | 61 | 68.9 | 63.7 | |
| Seedance 1.0Model Category=Commercial2026.06 | 55.1 | 54.2 | 42.5 | 44.8 | 45.4 | 44.2 | 62.2 | 64.1 | 69.8 | 68.6 | |
| Kling 2.6 ProModel Category=Commercial2026.06 | 53.4 | 52.9 | 59.8 | 36.4 | 53 | 35.8 | 57 | 60.5 | 63.7 | 61.3 | |
| Abot-PhysWorldModel Category=Robotics-specific2026.06 | 52.9 | 48.6 | 54.8 | 43.4 | 52.3 | 45.4 | 66.2 | 66.8 | 53.1 | 45.5 | |
| Wan2.2-I2V-A14BModel Category=Finetuned methods, Fine-tuning status=Original2026.06 | 50.7 | 38.1 | 45.4 | 37.3 | 50.1 | 33 | 60.8 | 58.2 | 69 | 64.8 | |
| Wan 2.2 A14BType=open-source2026.07 | 50.7 | 38.1 | 45.4 | 37.3 | 50.1 | 33 | 60.8 | 58.2 | 69 | 64.8 | |
| PhysisForcingModel Category=Finetuned methods, Backbone=Wan2.2-TI2V-5B, Fine-tuning status=ft2026.06 | 47.5 | 43.4 | 42.6 | 29.6 | 47.5 | 31.2 | 51.6 | 57.4 | 59.7 | 64.2 | |
| Cosmos 2.5Model Category=Robotics-specific2026.06 | 46.4 | 35.8 | 33.8 | 20.1 | 49.6 | 39.9 | 54.4 | 56 | 65.8 | 62.6 | |
| HunyuanVideo 1.5Model Category=Open-source2026.06 | 46 | 44.2 | 31.6 | 31.2 | 43.8 | 36.4 | 51.3 | 52.6 | 63.4 | 59.5 | |
| HunyuanVideo 1.5Type=open-source2026.07 | 46 | 44.2 | 31.6 | 31.2 | 43.8 | 36.4 | 51.3 | 52.6 | 63.4 | 59.5 | |
| Wan2.2-TI2V-5BModel Category=Finetuned methods, Fine-tuning status=ft2026.06 | 44.8 | 39.6 | 41.5 | 24.9 | 42.4 | 28.8 | 49.2 | 52.7 | 61.4 | 62.6 | |
| LongCat-VideoModel Category=Open-source2026.06 | 43.7 | 37.2 | 31 | 22 | 38.4 | 18.6 | 58.6 | 57.6 | 68.1 | 62.1 | |
| LongCat-VideoType=open-source2026.07 | 43.7 | 37.2 | 31 | 22 | 38.4 | 18.6 | 58.6 | 57.6 | 68.1 | 62.1 | |
| DreamGen(gr1)Model Category=Robotics-specific2026.06 | 42 | 31.2 | 37.2 | 29.7 | 33.4 | 21.5 | 56.4 | 53.2 | 57.9 | 57.5 | |
| DreamGen(droid)Model Category=Robotics-specific2026.06 | 40.5 | 35.8 | 34.8 | 21.4 | 31.6 | 33.9 | 49.9 | 47.6 | 54.2 | 55.6 | |
| Wan2.1-14BModel Category=Open-source2026.06 | 39.9 | 34.4 | 26.8 | 28.2 | 33.5 | 20.5 | 46.4 | 49.7 | 59.5 | 59.9 | |
| LTX-2Model Category=Open-source2026.06 | 38.1 | 28.4 | 30.4 | 23.3 | 38.6 | 16.4 | 45.3 | 42.4 | 62.2 | 55.5 | |
| Wan2.2-TI2V-5BModel Category=Open-source2026.06 | 38 | 33.1 | 31.3 | 14.2 | 31.8 | 23.4 | 43.6 | 44.8 | 59 | 60.7 | |
| Wan2.2-TI2V-5BModel Category=Finetuned methods, Fine-tuning status=Original2026.06 | 38 | 33.1 | 31.3 | 14.2 | 31.8 | 23.4 | 43.6 | 44.8 | 59 | 60.7 | |
| Sora v2 ProModel Category=Commercial2026.06 | 36.2 | 20.8 | 26.8 | 18.6 | 25.5 | 11.5 | 47.6 | 51.3 | 66.4 | 56.1 | |
| SkyReelsModel Category=Open-source2026.06 | 36.1 | 20.3 | 27.6 | 20.3 | 25.4 | 23.4 | 50.7 | 47.7 | 58.6 | 50.9 | |
| LTX-VideoModel Category=Open-source2026.06 | 34.4 | 30.2 | 17.6 | 21 | 28 | 24.1 | 44 | 45.6 | 52.6 | 46.4 | |
| FramePackModel Category=Open-source2026.06 | 33.9 | 20.6 | 25.8 | 17.3 | 16.9 | 17 | 44 | 46.4 | 62.6 | 54.8 | |
| HunyuanVideoModel Category=Open-source2026.06 | 30.3 | 17.7 | 18 | 10.8 | 14.7 | 3.5 | 45.4 | 48 | 62.5 | 52.4 | |
| Sora v1Model Category=Commercial2026.06 | 26.6 | 15.1 | 22.3 | 11.1 | 16.6 | 13.9 | 31.4 | 32.4 | 54.4 | 41.9 | |
| CogVideoX_5BModel Category=Open-source2026.06 | 25.6 | 11.6 | 11.2 | 9.8 | 21.2 | 7.9 | 33.8 | 38.5 | 46.5 | 49.6 | |
| VidarModel Category=Robotics-specific2026.06 | 20.6 | 7.3 | 10.6 | 5 | 5.4 | 5 | 38.2 | 41 | 37.4 | 35.7 | |
| UnifoLM-WMA-0Model Category=Robotics-specific2026.06 | 12.3 | 3.6 | 4 | 1.8 | 6.2 | 0 | 26.8 | 19.4 | 29.3 | 20 |