MiMo-Embodied 7B: When Robotics and Driving Finally Share a Brain

MiMo-Embodied 7B is Xiaomi’s open-source vision-language foundation model designed to work across both embodied AI (robots acting in the physical world) and autonomous driving. Instead of training separate models for robots and vehicles, Xiaomi unified them into one system, and surprisingly, it works better that way. This makes MiMo-Embodied 7B the first open VLM that treats indoor embodied reasoning and outdoor driving perception as related problems rather than isolated domains.

One Model, Two Worlds