Tools
Robot policies can move but can't think. LLMs can think but can't move. So we connected them. Real robot: 16.7% → 97.3% Sim (LIBERO-PRO): 12…
A team led by Liane Galanti linked large‑language models (LLMs) with robotic motion policies, allowing a robot to combine reasoning capabilities with physical movement. In experiments on a real robot,
A team led by Liane Galanti linked large‑language models (LLMs) with robotic motion policies, allowing a robot to combine reasoning capabilities with physical movement. In experiments on a real robot, task success rose from 16.7 % pre‑integration to 97.3 % after the connection, while in LIBERO‑PRO simulation performance improved from 12.8 % to 53.3 %. This demonstrates that integrating LLMs can bridge the traditional divide between “moving” and “thinking” in robotics.
Source: Together AI (X) | 2026-07-28