← sources

“Can Qwen directly control the robot as well as GPT-6 Astra?”

source: reddit (reddit:LocalLLaMA)published 2026-11-03 17:03ingested 2026-11-04 02:41

retained text

A open source harness that was used with Qwen3.8 FN that QwenDev did a shout out to. Available now.

From QwenDev via x

https://x.com/QwenDevs/status/2117644199860601297

“Well, turns out Qwen3.8-Flash-Next can do more than work with software tools, it can operate real robots too.

A group of Qwen developers from Tsinghua and UC Berkeley recently built uni-robot-agent, an open-source harness that connects general-purpose LLMs to robotic systems. It gives Qwen a way to interact with the physical world through tool calls without changing the model weights.

With this harness, Qwen3.8-Flash-Next reached the performance level of some previously reported closed-model robotic agent systems. The average API cost was $1.09 per task, compared with $21.29 for GPT-6 Astra. On 19 real-world robot manipulation tasks. Without any robot-specific fine-tuning, Qwen achieved a slightly higher score than the fine-tuned π₀.₅

The harness is open-source if anyone wants to tinker with it.”

From:Yuxin Zuo via X

[https://x.com/zuo\_yuxin/status/2117635294115795134](https://x.com/zuo_yuxin/status/2117635294115795134)

“Introducing uni-robot-agent: a unified harness for Robot-Use Agents (RUAs), a first step toward RUA training and local deployment.

We tested uni-robot-agent with Qwen3.8-Flash-Next running locally:  
 On s, its score rose from 1.90 to 31.39, surpassing the RoboDojo Team’s score with GPT-6 Astra.  
 On real-world tasks, it outscored fine-tuned π0.5 by 5.75 points and zero-shot Dream-1 by 10 npoints.”

document fetched 2026-11-04 00:41 · canonical https://reddit.com/r/LocalLLaMA/comments/1xmyji8/can_qwen_directly_control_the_robot_as_well_as