China’s Wukong AI, named after the mythic Monkey King, is now the nation’s first large language model (LLM) deployed aboard its Tiangong space station, marking a milestone in integrating AI systems into crewed spaceflight, says Wired.com.
Developed from a domestic open-source model, Wukong is tailored to the demands of human space missions. It features a dual-module design: one resides aboard the station for instant, in-orbit support, while the other on Earth handles more complex analysis and adaptive planning. This hybrid architecture enhances responsiveness and depth of assistance.
Wukong’s debut coincided with a 6.5-hour spacewalk, supporting taikonauts in installing debris shields and performing routine inspections. The AI provided real-time guidance, helped with mission coordination, and bolstered psychological support—a notable benefit during intense EVA operations.
While not the first AI in space—catching up to systems such as CIMON and Astrobee aboard the ISS—Wukong distinguishes itself through its focus on space navigation and mission orchestration, offering a more multi-functional AI assistant rather than a single-use robotic or conversational aid.
The deployment of Wukong reflects broader ambitions: Tiangong is central to China’s long-term vision as a lunar logistics hub and training platform for deep-space missions. Embedding AI like Wukong in this environment boosts operational efficiency and paves the way for more autonomous and distributed space station capabilities.
Wukong AI exemplifies the evolution of AI systems in space, providing real-time navigation aid, mission planning, and psychological backup aboard Tiangong. Its hybrid Earth–space architecture enhances adaptability and underscores China’s push toward autonomous, intelligent orbital operations.