NVIDIA Cosmos models have reached 10 million downloads on Hugging Face. This milestone belongs to the developers and researchers building with open world foundation models for physical AI.
Cosmos 3 is the world’s first fully open omnimodel, unifying vision reasoning, world generation and action generation.
What can you do with Cosmos?
- Reason over live video to generate insights and alerts.
- Simulate physical environments and predict future world states for training and evaluation.
- Generate synthetic data across text, images, video, ambient sound and actions.
- Post-train specialized world action models for robots, autonomous vehicles and other physical AI systems.
Cosmos 3 models, the post-training framework, data-processing libraries, datasets and agentic skills are open—giving developers control to customize, deploy and optimize their models.
The newest addition, Cosmos 3 Edge, brings an open 4B world model to edge devices for real-time reasoning and robot-action generation.
To everyone who downloaded a model, tested an idea, shared feedback or built with Cosmos: thank you for being part of this milestone. 💚
Explore NVIDIA Cosmos on Hugging Face