On October 15, Xiaomi announced a fully open source MidashenGLM-Spatial. According to Xiaomi, this model is the industry's first open source end-to-end spatial audio understanding model. While retaining basic audio comprehension capabilities, it has spatial perception and reasoning capabilities. Xiaomi said that the model will be successively applied to scenarios such as Xiaomi smart homes and car cockpits to further enhance the interactive experience of “the entire ecosystem of people, cars, and households” with spatial hearing ability.

On October 15, Xiaomi announced a fully open source MidashenGLM-Spatial. According to Xiaomi, this model is the industry's first open source end-to-end spatial audio understanding model. While retaining basic audio comprehension capabilities, it has spatial perception and reasoning capabilities. Xiaomi said that the model will be successively applied to scenarios such as Xiaomi smart homes and car cockpits to further enhance the interactive experience of “the entire ecosystem of people, cars, and households” with spatial hearing ability.