NVIDIA released: Cosmos Reason 2, the latest reasoning visual language model, specializing in physical AI, with a context length of 256K
NVIDIA has launched Cosmos Reason 2, the latest inference visual language model focused on physical AI, supporting context lengths up to 256K. This model improves spatiotemporal understanding and timestamp accuracy, is capable of 2D/3D point positioning, bounding box coordinates, trajectory data and OCR, and outputs robot actions and motion trajectories. It is suitable for applications such as video analysis, data annotation and safety detection. Available in 2B and 8B models
NVIDIA has launched Cosmos Reason 2, the latest inference visual language model focused on physical AI, supporting context lengths up to 256K. This model improves spatiotemporal understanding and timestamp accuracy, is capable of 2D/3D point positioning, bounding box coordinates, trajectory data and OCR, and outputs robot actions and motion trajectories. It is suitable for applications such as video analysis, data annotation and safety detection. Available in 2B and 8B models