Xiaomi-Robotics-U0

Xiaomi-Robotics-U0 is a 38-billion parameter multimodal autoregressive foundation model for embodied AI developed by Xiaomi Robotics team. It unifies multiple capabilities including embodied scene generation from text descriptions, trajectory transfer to new scenes, robot interaction video generation, and general image generation and editing. The model generates robot-ready environments directly from text, serving as a versatile foundation for embodied AI applications that bridge language, vision, and physical action understanding.