Microsoft has introduced three AI audio models for conversational applications: MAI-Transcribe-2-Streaming, MAI-Voice-2.1 and MAI-Voice-2.1-Flash. The October 1 launch combines live speech recognition with multilingual voice generation for developers building assistants and other interactive services.
The transcription model handles 60 languages and detects language changes automatically. It returns preliminary text as speech arrives, allowing applications to begin processing a request before the speaker finishes. Introductory pricing is $0.54 per audio hour through year-end.
The two voice models support 23 languages and 26 locales. MAI-Voice-2.1 costs $22 per million characters, while the Flash variant costs $15 and is designed for applications that prioritize response speed and high usage volumes.
All three are available through Microsoft Foundry, with the voice models also offered through OpenRouter. Microsoft says voice cloning includes consent safeguards. Developers can combine recognition and generation with their own reasoning and tool-use systems.
The release follows Google’s Gemini Live launch, which emphasized uninterrupted dialogue and background tool calls. Together, the launches show suppliers competing over the components needed for voice agents that can complete tasks while keeping a conversation moving.
Disclaimer: AIstify is an independent media brand owned and operated by NuvexMedia LLC, publishing news, research, and insights on artificial intelligence, emerging technologies, automation, and related industries. NuvexMedia LLC invests in and collaborates with companies across the AI, technology, software, and digital innovation sectors. These relationships do not influence AIstify’s editorial coverage, and the publication maintains full editorial independence to provide accurate, timely, and objective information. © 2026 NuvexMedia LLC. All rights reserved. This content is for informational purposes only and should not be considered legal, tax, investment, financial, or other professional advice.
