uni-robot-agent brings open-weight models to robot control
Researchers from Tsinghua and UC Berkeley introduced uni-robot-agent, an open-source harness that connects general-purpose LLMs to robotic systems through tool calls without changing model weights. The paper reports that Qwen3.8-Flash-Next scored 31.39 on RoboDojo without privileged inputs, above the published 28.97 for GPT-6 Astra, and performed competitively across 19 real-world tasks. These results are reported by the paper authors and Qwen developers and have not been independently verified here.
The harness is designed to make open-weight robot-use agents easier to train, deploy locally and study without relying on frontier proprietary models.
Computer Science > Robotics
arXiv:2611.01560 (cs)
[Submitted on 1 Nov 2026]
Title:Democratizing Robot-Use Agents: A Unified Harness with Open Models
Authors:Jiarui Yuan, Yuxin Zuo, Xiaoyu Huang, Wenbo Zhang, Kwanyoung Park, Junlin Yang, Tianxiang Zhang, Ruoqu Chen, Yutao Ouyang, Bingxiang He, Xingtai Lv, Hongjin Wu, Jiajun Liu, Jiahao Zhao, Tianwei Luo, Haozhan Li, Cheng Qian, Jiaxin Ge, Huan-ang Gao, Kaisen Yang, Jiaqi Deng, Qixin Xu, Siqiao Huang, Bowen Zhou, Mengdi Xu, Zhiyuan Liu, Pieter Abbeel, Ning Ding
View a PDF of the paper titled Democratizing Robot-Use Agents: A Unified Harness with Open Models, by Jiarui Yuan and 27 other authors
#uni-robot-agent#Qwen3.8-Flash-Next#GPT-6 Astra#RoboDojo#π0.5#Dream-1#Tsinghua University#UC Berkeley