NVIDIA has released Cosmos 3 Edge, a 4-billion-parameter open world model, on Hugging Face’s Cosmos 3 repository. The model is intended to help robots and vision AI agents understand their surroundings, reason in real time, and generate robot actions on edge devices. It is optimized for memory efficiency and high-throughput inference across NVIDIA edge computing platforms, including the newly announced Jetson T2000 and T3000 modules. The model is designed to serve as a compact vision language model (VLM) with best-in-class throughput and accuracy for real-time inference. Source: huggingface
Cosmos 3 Edge operates at robot-control resolution (640×360 observations) and delivers real-time reasoning, generating 32 actions per inference on NVIDIA Jetson Thor while achieving real-time control at 15 Hz. Among similar size (4B parameters) models, Cosmos 3 Edge ranks #1 on VANTAGE-Bench for vision analytics and is state-of-the-art for robot policy learning. The model sets the state of the art for smart infrastructure and robotics. Source: huggingface
World modeling involves learning how an environment changes over time, representing objects, motion, spatial relationships, and the effects of actions. A world model helps robots reason about these relationships, predict the visual result of an action, infer the action that caused a change, or generate an action to produce a desired outcome. Cosmos 3 Edge brings these capabilities together in one model, on-device. Its shared representation enables physical AI systems to understand the current world state, simulate possible futures, and connect those futures to actions. Source: huggingface
