Incentivizing Multimodal Reasoning in Large Models for Direct Robot Manipulation

Open in new window