NVIDIA this week launched NVIDIA Cosmos 3, an open world foundation model for physical AI built on a mixture-of-transformers architecture that combines vision reasoning, world generation and action prediction in a single system.
Cosmos 3 is the world’s first fully open omnimodel that can natively understand and generate text, images, video, ambient sound and actions with leading physics accuracy, reducing physical AI training and evaluation cycles from months to days.
NVIDIA also launched the NVIDIA Cosmos Coalition, a global collaboration between world model builders and AI developers — including Agile Robots, Black Forest Labs, Generalist, LTX, Runway and Skild AI — working together to advance next-generation world models.
“The big bang of physical AI is just around the corner thanks to breakthroughs in multimodal reasoning language, vision and world models,” says Jensen Huang, founder and CEO of NVIDIA. “The Cosmos 3 family of open, frontier omnimodels gives developers a generational leap in ability to build robots, autonomous vehicles and vision AI that perceive, reason, plan and act in the physical world.”
A new architecture for physical AI
Cosmos 3 tackles a fundamental challenge in physical AI: enabling robots, autonomous vehicles (AVs) or vision agents to generalize in the real world with limited training data and fragmented simulation stacks.
The model’s mixture-of-transformers architecture pairs a reasoning transformer with an expert generation transformer, enabling Cosmos 3 to understand object interactions, motion and spatial-temporal relationships before generating video and action trajectories.
Trained on one of the largest multimodal physical AI datasets — including billions of samples across text, image, video, sound and action trajectories — the model gives developers a powerful pretrained foundation for building physical AI systems with less data and lower training costs.
Developers can use Cosmos 3 as:
- A vision language model that understands and reasons across modalities.
- A world model or video foundation model that simulates physical environments and predicts future world states for training and evaluation.
- The backbone for world action models that help train robots to perform specific tasks.
Cosmos 3 models deliver leading results on physical AI benchmarks. Among open models, it ranks first across Artificial Analysis, Physics-IQ, PAI-Bench and R-Bench for world generation accuracy, RoboLab and RoboArena for action policy, and the VANTAGE-Bench and TAR leaderboards for vision understanding.
The Cosmos 3 lineup gives developers options for different stages of physical AI development:
- Cosmos 3 Super for post-training robotics and AV models that need the highest physics accuracy and generation quality.
- Cosmos 3 Nano for high-quality video and action reasoning in fractions of a second.
- Cosmos 3 Edge, coming soon, for real-time inference at the edge.