We’re thrilled to announce native support for NVIDIA’s powerful new model suite — Cosmos-Predict2 — in ComfyUI!
Model Overview
Cosmos-Predict2 is a family of highly performant pre-trained world foundation models developed by NVIDIA, purpose-built for generating physics-aware images, videos, and world states for Physical AI development.
The models are available for commercial use under the NVIDIA Open Model License.
You can find all the supported models in Comfy-Org/Cosmos_Predict2_repackaged
Get Started
Update ComfyUI to the latest
Go to
Workflow → Template
, and find the Cosmos templates, or download the example workflows belowDownload the models as instructed and run!
Cosmos-Predict2 Text-to-Image Workflow
Example Outputs
Check our documentation for more details.
Cosmos-Predict2 Video2World (Image-to-Video) Workflow
Example Outputs
Check our documentation for more details and enjoy creating!