Alibaba (09988) officially launched the Qwen-Robot series of embodied intelligence large models on June 16, which includes three models: the Vision-Language-Action manipulation model Qwen-RobotManip, the Vision-Language-Navigation mobility model Qwen-RobotNav, and the world model Qwen-RobotWorld. Qwen-RobotNav serves as the action gateway for physical agents, unifying four task types: instruction following, goal navigation, object tracking, and autonomous driving. Qwen-RobotManip forms the interaction foundation for physical agents, achieving large-scale multi-robot training based on a corpus of over 38,100 hours built entirely from open-source data. Qwen-RobotWorld focuses on the infinite world of physical agents, enabling a single world model to forecast physically plausible futures across manipulation, driving, and navigation scenarios. All three models provide language-first interfaces and can be compositionally invoked via the general-purpose Qwen models. Alibaba also disclosed its internal robotic agent framework, Qwen-RobotClaw, which enables Qwen VLM agents to invoke the Qwen-Robot Suite models as tools for interacting with the physical world. Additionally, on May 20, Qwen officially released Qwen3.7-Max, a next-generation flagship model designed for the era of intelligent agents, which will soon be available via API.