88Z AI, 16 June — Alibaba Tongyi has released the Qwen-Robot Suite, integrating navigation, manipulation, and world prediction models. It aligns vision and action via a language-first interface, achieving SOTA results in over ten benchmarks.
At a glance
- The suite includes navigation, manipulation, and world prediction models, aligning vision and action via a language-first interface.
- It achieves SOTA performance in over ten benchmarks covering navigation, manipulation, and world modeling.
The update
Alibaba Tongyi has officially launched the Qwen-Robot Suite, a foundation model kit designed for the physical world. The suite features three core models: RobotNav for mobile path planning, RobotManip for precise action execution, and RobotWorld for environmental dynamic deduction. By adopting a language-first interface design, the kit effectively aligns the understanding capabilities of vision-language models with specific physical action commands. This architecture significantly enhances interaction and execution efficiency for robots in complex environments. Currently, the suite has achieved state-of-the-art performance levels in more than ten authoritative benchmarks covering navigation, manipulation skills, and world model prediction, demonstrating its strong potential in the field of embodied intelligence.
Sources & editorial note
This article is an AI-assisted synthesis by 88Z AI based on the public sources below. Product claims and availability are attributed to their publishers. It does not represent independent testing or first-hand reporting.
- 千问大模型Source published
Source dates above refer to the original announcements; the date under the headline is this article’s publication on 88Z AI.