NVIDIA Releases Cosmos-Reason2-32B Flagship Model Weights, Expands Context Window to 256K Tokens

According to Beating, NVIDIA has released the weights for Cosmos-Reason2-32B, the flagship version of its physical AI reasoning vision-language model (VLM) designed to help robots and autonomous driving systems understand spatial, temporal, and physical principles. The 32-billion-parameter model, built on Qwen3-VL-32B-Instruct, is available under NVIDIA Open Model License for commercial use.

The model can analyze video streams to assess driving safety, detect objects in images with 2D/3D coordinates, and serve as a planning engine for humanoid robots and autonomous vehicles. Compared to the previous generation, Cosmos-Reason2 adds object detection with precise temporal localization and expands the context window to 256K tokens.