NVIDIA released Cosmos 3, an open-weights physical AI foundation model designed to enable teams to build specialized sys
In July, NVIDIA joined over 200 companies and organizations in signing an open letter titled "Open Weights and American AI Leadership," which argues that AI leadership will be measured by whether an open ecosystem reaches every sector, not by any single frontier model. Open models that anyone can download, inspect, modify and run on their own infrastructure are essential to making this possible, especially in physical AI where every deployment requires specialization.
Physical AI systems must understand and predict consequences, not just appearances. World models learn how physical environments behave, what may happen next and which actions make sense. They generate physically grounded world and action data, simulate future states and provide a foundation that teams can specialize for robots, autonomous vehicles or vision AI systems. Open world models are already being used to generate training data, test policies and specialize physical AI systems. NVIDIA Cosmos 3 brings these capabilities together in an open model family with leading benchmark results and adoption across robotics, autonomous vehicles and vision AI.
The data required for physical AI is difficult and expensive to collect at the necessary scale. Collecting rare events and long-tail scenarios safely and repeatedly is especially challenging. Open world models help by providing more useful data through learning physical relationships from large-scale multimodal scenarios, more diverse environments that vary in weather, lighting, objects and trajectories, and a better foundation to build on and adapt to specific robots, vehicles, sensor configurations, tasks or operating environments.
Teams need access to model weights, a license permitting adaptation and the tools needed for post-training to close the gap between a general model and their particular systems. NVIDIA Cosmos world foundation models are available under the Linux Foundation's OpenMDW 1.1 license, enabling teams to post-train models on their own data and hardware. Specializing a model requires not only the model itself but also environments to generate data, run simulations and test behavior. Omniverse libraries help developers build simulation-ready environments, while OpenUSD provides the framework for composing, reusing and exchanging complex 3D data across digital twins, simulations and synthetic data generation workflows.
NVIDIA Cosmos 3 is a frontier open physical AI foundation model built on a mixture-of-transformers architecture that combines vision reasoning, world generation and action prediction. This allows developers to use one model family to understand scenes, generate synthetic data, simulate future states and build specialized world action models, rather than assembling and maintaining separate models for each capability. The family includes Cosmos 3 Super at 64 billion parameters for high-fidelity world modeling, Cosmos 3 Nano at 16 billion parameters for efficient reasoning and post-training, and Cosmos 3 Edge at 4 billion parameters for on-device vision reasoning and robot policy deployment. Cosmos 3 Edge is lightweight enough to run on edge GPUs, NVIDIA RTX GPUs, NVIDIA DGX systems and NVIDIA Jetson platforms, including Jetson Thor.
Cosmos 3 ranks first on Artificial Analysis for open weights text-to-image and image-to-video generation, on PAI-Bench for world generation, in the image-to-video category of Physics-IQ benchmark, and on RoboLab for robot policy. Cosmos 3 Super is also the highest-ranked open model on VANTAGE-Bench for vision understanding. NVIDIA's broader physical AI stack includes Isaac GR00T for robotics, Alpamayo for autonomous vehicles and Metropolis for vision AI.
Major companies are building on NVIDIA Cosmos for physical AI applications, including Doosan Robotics, LG Electronics, Samsung Electronics and Skild AI in robotics; Li Auto, Xiaomi and Afari in autonomous vehicles; and Centific, Fogsphere, Linker Vision, Milestone Systems and Yuan for vision AI agents in industrial AI and smart spaces. The NVIDIA Cosmos Coalition brings together world model builders, AI developers and physical AI leaders to contribute models, research and evaluation methods. NVIDIA recently expanded the coalition to Japan, where robotics and manufacturing leaders plan to join and develop open world models for factories, logistics, agriculture, construction, healthcare and transportation.