MiMo-V2.5-ASR is a powerful 8-billion parameter open-source speech recognition model developed by Xiaomi, engineered for exceptional accuracy and versatility. This advanced model proficiently transcribes a wide array of languages and speech styles, including Mandarin, English, eight distinct Chinese dialects, complex code-switched speech, and even song lyrics. It is an indispensable tool for machine learning engineers, researchers, and developers focused on creating robust, real-world voice applications that demand high-performance and broad linguistic coverage.
Categories:
API
Launch Date:
April 28, 2026