Microsoft Launches New Streaming Transcription and Voice AI Models

Microsoft AI has announced the release of new streaming transcription and voice models, including MAI-Transcribe-2-Streaming, MAI-Voice-2.1, and MAI-Voice-2.1-Flash. MAI-Transcribe-2-Streaming has debuted at the top of Artificial Analysis's streaming speech-to-text leaderboard. The new MAI-Voice-2.1 model offers enhanced multilingual text-to-speech capabilities, supporting 23 languages and 26 locales with native accents, while MAI-Voice-2.1-Flash is designed for high-volume applications. These models aim to provide faster and more fluid building blocks for conversational AI experiences without compromising accuracy or voice quality.