Xiaomi Open-Sources MiDashengLM-Spatial, Claimed as First End-to-End Spatial Audio Understanding Model

On October 15, Xiaomi announced the full open-source release of MiDashengLM-Spatial, which it described as the industry's first open-source end-to-end spatial audio understanding large model. The model retains basic audio understanding capabilities while adding spatial perception and reasoning abilities. Xiaomi said the model will be gradually applied to scenarios such as smart home and automotive cockpits, aiming to enhance the interaction experience of its "Human x Car x Home" ecosystem through spatial hearing.