All tasks / Models
Cosmos3-Super-Image2Video
NVIDIA's open-weights 64B omnimodal world model; this checkpoint turns an image plus a text prompt into video.
- NVIDIA
- Open weights
- Released 31 May 2026
Where it ranks
- AVideo generationModels · Local · 1,275 on Artificial Analysis Video Arena: Image to Video, no audio (AA-Video-I2V-Silent v1.0) (secondary benchmark)
Facts
- Vendor
- NVIDIA
- Family
- Cosmos
- Released
- 31 May 2026
- Weights
- Open: download and run it yourself
- License
- OpenMDW 1.1
- Parameters
- 64B
- Access
- Free (open weights)
- Last verified
- 19 Oct 2026
Price from huggingface.co, 19 Oct 2026.
Benchmark results
| Benchmark | Result | Entry on the leaderboard | As of | Source |
|---|---|---|---|---|
| Artificial Analysis Video Arena: Image to Video, no audio (AA-Video-I2V-Silent v1.0) | 1,275 | Cosmos3-Super-Image2Video-4StepAA-Video-I2V-Silent v1.0 (no audio) board; 4-step distilled variant (base checkpoint scores 1252.79) | 19 Oct 2026 | artificialanalysis.ai |
Where you can use it
No tracked tool lists this model yet; as open weights it runs in local runtimes.