MAI-Voice-2.1

MAI-Voice-2.1 is a text-to-speech model released on 1 October 2026 by Microsoft AI. It supports 23 languages and 26 locales and lets a single voice speak all of them with a native accent, so a brand or assistant can keep one voice while switching language mid-conversation. It can clone a voice across all supported languages from a few seconds of reference audio, and Microsoft says built-in consent guardrails are meant to prevent misuse. Microsoft prices it at $22 per million characters. It is available through Microsoft Foundry, the MAI Playground, Vercel, Azure Voice Live and OpenRouter. A faster, cheaper variant, MAI-Voice-2.1-Flash, was released the same day for high-volume use. The weights are not published, and the quality claims are Microsoft’s own.

Not yet independently verified: Every detail here comes from Microsoft AI’s own announcement of 1 October 2026. The accuracy ranking and speed comparisons are Microsoft’s statements; we have not read the ranking site’s own page, and no independent press report had been found at 2026-10-04. We will update this when it can be confirmed, and remove this note.