Model
CogAgent
A vision-language model trained specifically for GUI operation tasks. It interprets screenshots and executes desktop automation actions, supporting visual grounding and interaction with UI elements across graphical interfaces.
Timeline (last 90 days)
Aug 26
基于视觉理解的移动端自动化方案代表包括Google DeepMind的CogAgent、微软的UFO项目以及Apple Intelligence中的App Intents框架
Unverified50%