Tsinghua and Berkeley open-source uni-robot-agent, Qwen beats GPT-6 zero-shot

Tsinghua and UC Berkeley released uni-robot-agent, an open-source framework letting open-weight LLMs control real and simulated robots via tool calls, with Qwen zero-shot outperforming GPT-6 Astra's public scores at roughly 6.6% of the cost.

2026-11-03 ~ 2026-11-04 · 4 related posts

• Tsinghua & Berkeley open-source uni-robot-agent: Qwen beats GPT-6 zero-shot — Jiarui Yuan · 2026-11-03
• Tsinghua & Berkeley open-source uni-robot-agent: Qwen drives real robots at $1.09/task — Beneficial-Ad-8127 · 2026-11-04
• Open-weight LLMs control robots zero-shot: Qwen scores 31.39 on RoboDojo, beating GPT-6 Astra — pabbeel · 2026-11-04

1 near-duplicate retellings: 青稞AI