NVIDIA Adds Four Open Cosmos 3 Simulation Checkpoints
In a company blog post, NVIDIA released four open-weight checkpoints for its Cosmos 3 world model platform. Cosmos3-Nano-Transfer-Auto, Cosmos3-Nano-Sim-Auto, Cosmos3-Nano-Sim-Transfer, and Cosmos3-Nano-Sim-Bimanual each have 16 billion trainable parameters and use the OpenMDW 1.1 license for commercial and noncommercial use.
Cosmos3-Nano-Transfer-Auto generates synchronized driving video across as many as 11 camera views, with optional joint RGB and LiDAR output. Developers can control scene structure, traffic, weather, lighting, and time of day. Cosmos3-Nano-Sim-Auto is a distilled streaming version that generates successive camera and LiDAR segments from driving controls and previous output.
Cosmos3-Nano-Sim-Transfer converts simulator controls and text prompts into video while leaving physics and scene state calculations to the simulator. Cosmos3-Nano-Sim-Bimanual predicts observations from camera movements and two-arm robot actions, with support for YAM and AgiBot systems.
Running the published setup requires Linux, an NVIDIA GPU, Docker, and NVIDIA Container Toolkit.
We hope you enjoyed this article
Consider subscribing to one of our newsletters like Robotics Brief or Daily AI Brief.
Also, consider following us on social media:
More from Robotics
Nov 5 Serve Robotics Launches Physical AI Services for Robot Makers Nov 5 Axon Invests $1.2 Billion in Skydio Nov 5 Agility Robotics Shares Slide in Nasdaq Debut Nov 5 Zipline Raises $1.5 Billion at $21.5 Billion Valuation Nov 5 Humble Robotics Raises $51 Million for Cabless Autonomous TrucksRobotics Brief
Weekly coverage of AI-driven robotics advances in industrial automation, autonomous vehicles, and robotic systems.
Industry analysis
2025 Global Business Services Agenda: Gen AI Takes Center Stage
This industry analysis by The Hackett Group explores the transformative impact of generative artificial intelligence (Gen AI) on global business services (GBS) in 2025. The study highlights the shift from exploration to acceleration of Gen AI initiatives, with 89% of executives advancing these projects to improve customer satisfaction, innovate products, and reduce costs. The report also discusses the challenges and strategies for successful Gen AI adoption, emphasizing the need for a technology-enabled operating model and the importance of reskilling the workforce.
Read moreYou may also like
Hesai Launches Kosmo 3D Data Engine for Robot Training
Volkswagen Builds Autonomous Vehicle Data System With NVIDIA, Microsoft and Voxel51
NVIDIA Expands Isaac With Agile ONE SN Humanoid and Robot Learning Tools
Nebius Releases Physical AI Workbench for Robotics
NVIDIA Expands Hyperion Autonomous Vehicle Platform in Europe
Daily AI Brief: the AI news that matters, in your inbox.