Alibaba Qwen Team Releases Mobile-Agent-v3 and GUI-Owl: Next-Generation Multi-Agent Framework for GUI Automation

Alibaba's Qwen team has developed Mobile-Agent-v3 and GUI-Owl, a next-generation multi-agent framework designed to automate graphical user interface (GUI) tasks across mobile, desktop, and web platforms. These models leverage advanced vision-language capabilities, with GUI-Owl built on the Qwen2.5-VL foundation and trained on extensive GUI interaction datasets, enabling it to understand screens, reason about tasks, and execute actions in a human-like manner. The key innovation lies in GUI-Owl's unified, end-to-end multimodal architecture that integrates perception, grounding, reasoning, planning, and action