Reproduce “Democratizing Robot-Use Agents: A Unified Harness with Open Models” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Jiarui Yuan, Yuxin Zuo, Xiaoyu Huang, Wenbo Zhang, Kwanyoung Park, Junlin Yang, Tianxiang Zhang, Ruoqu Chen and 20 more
The authors introduce uni-robot-agent, a shared robot-interaction harness that lets adapters translate common tools and execution feedback across models and environments. On RoboDojo, open-weight Qwen3.8-Flash-Next scores 31.39 without privileged simulator inputs, above the published GPT-6 Astra–RoboProbe score of 28.97; this is not a matched comparison, since evaluations used 252 and 2,100 episodes, respectively.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.