Skip to content
NewA calmer, faster ZeroHour

OpenRUA: Robot-Use Agents Are Zero-Shot Visuomotor Policies

Hugging Face daily papersby Zhaoyang Chu, Earl T. Barr, Claire Le Goues ()1 min read2 · 1 read
AI summary · grok-4.7Summaries by Vani AI

OpenRUA lets Claude Opus 5 control robots through native ROS 2, scoring 99% on CaP-Bench without custom harnesses.

OpenRUA is a minimal harness that gives off-the-shelf coding agents only terminal access to a robot's native ROS 2 interface, without bespoke primitives or prescribed workflows. Using Claude Code with Claude Opus 5, it reports 99.0% success on CaP-Bench and 87.0% on LIBERO-PRO as a zero-shot visuomotor policy. The agent spontaneously wrote perception programs in 96.80% of episodes, motion-control clients in 95.87%, and closed-loop controllers in 50.13%. Code is released at github.com/terminalworld/OpenRUA.

Key facts
  • Zero-abstraction harness exposes only ROS 2 docs, tools, and a terminal.
  • Claude Opus 5 scores 99.0% on CaP-Bench and 87.0% on LIBERO-PRO.
  • Agents derive metric perception from raw sensors in 96.80% of episodes.
  • Closed-loop control emerges in 50.13% of manipulation episodes.
Explain this AI news with Vani AIWhat was announced, what is new and why it matters, in a new chat

From the original article

1 min read at huggingface.co

Coding agents are extending their reach into the physical world by writing and executing robot control programs. One might expect the agents to use the existing mature software stack that engineers have developed over decades to access sensors and control motion.

Hugging Face daily papers

By Zhaoyang Chu, Earl T. Barr, Claire Le Goues · huggingface.co

Read the full article at Hugging Face daily papers

A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.