Model release · Microsoft
MAI-Voice-2.1 and Flash: two multilingual TTS variants
- Updated
- Released
Both TTS variants support 23 languages and preserve speaker identity across languages. Standard 2.1 targets high-quality, long-form audio at $22 per million characters; Flash targets lower latency and higher volume at $15. Microsoft reports 45 seconds of audio generated with 150 ms end-to-end latency for Flash; this is a vendor claim. Official access paths include Foundry and OpenRouter, with consent and access requirements for voice cloning.
Availability
Official access paths include Foundry, MAI Playground, Vercel, Azure Voice Live and OpenRouter. Voice cloning is subject to consent and access requirements.