Cosmos

Cosmos is a world model development platform that consists of world foundation models, tokenizers and video processing pipeline to accelerate the development of Physical AI at Robotics & AV labs. Cosmos is purpose built for physical AI. The Cosmos repository will enable end users to run the Cosmos models, run inference scripts and generate videos.

7619
489
Python

NVIDIA Cosmos Header

Thank you all for the valuable feedback! We have restructured the codebase to make it easier to use and contribute to.

New GitHub page for NVIDIA Cosmos: https://github.com/nvidia-cosmos

NVIDIA Cosmos now includes three subprojects:

  1. Cosmos-Predict1 is a collection of general-purpose world foundation models for Physical AI that can be fine-tuned into customized world models for downstream applications.
  2. Cosmos-Transfer1 is a world-to-world transfer model designed to bridge the perceptual divide between simulated and real-world environments.
  3. Cosmos-Reason1 models understand the physical common sense and generate appropriate embodied decisions in natural language through long chain-of-thought reasoning processes.

This repository is no longer maintained and will soon be deprecated. To view the initial release of NVIDIA Cosmos from this repository, please check out branch archived-ces2025.