# MAI-Voice-2.1-Flash

> Microsoft AI's low-latency text-to-speech model, announced 1 October 2026 — 45 seconds of audio with 150ms end-to-end latency, in 23 languages, for voice agents.

- Maker: Microsoft
- Line: MAI-Voice (text-to-speech)
- Canonical page: https://ai-tldr.dev/models/mai-voice-2-1-flash/
- Updated: 2026-10-04

## In short

MAI-Voice-2.1-Flash is the fast variant of MAI-Voice-2.1, announced by Microsoft AI on 1 October 2026 together with MAI-Voice-2.1 and the streaming speech-to-text model MAI-Transcribe-2-Streaming. Microsoft says it "supports the same languages, and cross-language speakers, as MAI-Voice-2.1" — 23 languages and 26 locales — "but it's been leveled up for high-volume, latency-sensitive workloads.". You can use it through Microsoft Foundry (Azure Speech), OpenRouter and Vercel AI Gateway.

## Facts

| Fact | Value | Source |
| --- | --- | --- |
| Input price | $15 / 1M characters | [1] |
| Licence | Proprietary · API only | [2] |
| Released | 1 October 2026 | [2] |

## Sources

1. [Microsoft AI — Our first streaming transcription model debuts at no. 1 on Artificial Analysis (MAI-Voice-2.1-Flash announcement)](https://microsoft.ai/news/our-first-streaming-transcription-model/)
2. [Microsoft Learn — MAI-Voice-2.1 and MAI-Voice-2.1-Flash](https://learn.microsoft.com/en-us/azure/ai-services/speech-service/mai-voices)

## Overview

MAI-Voice-2.1-Flash is the fast variant of MAI-Voice-2.1, announced by Microsoft AI on 1 October 2026 together with MAI-Voice-2.1 and the streaming speech-to-text model MAI-Transcribe-2-Streaming. Microsoft says it "supports the same languages, and cross-language speakers, as MAI-Voice-2.1" — 23 languages and 26 locales — "but it's been leveled up for high-volume, latency-sensitive workloads."

The launch post puts numbers on that: the model "can generate 45s of audio, with an end-to-end latency, of a mere 150ms", and Microsoft says it delivers 55% faster model inference and is about 60% cheaper than comparable models. Microsoft Learn positions it for real-time voice agents and assistants, low-latency call-centre and IVR flows, and multilingual interactive experiences.

Microsoft pitches it as the speaking half of a voice agent: "Pairing MAI-Transcribe-2-Streaming with MAI-Voice-2.1-Flash buys time back on both ends" — the streaming model transcribes the caller quickly and Flash answers at conversational speed. Like MAI-Voice-2.1, it supports fine-grained emotion and style control through SSML `mstts:express-as`, a library of licensed prebuilt voices, and gated instant voice cloning from a 5–60 second consented reference clip.

On Microsoft Foundry the model is in public preview, served globally from 14 Azure regions. It is called through the Azure Speech SDK with SSML — a prebuilt voice takes the model as a suffix, such as `en-US-Harper:MAI-Voice-2.1-Flash` — and through Azure Voice Live. Microsoft's launch post lists a price of $15 per 1M characters and access through Microsoft Foundry, MAI Playground, OpenRouter, Vercel and Azure Voice Live, with LiveKit support marked as coming soon.

## Pricing

- Input: $15 / 1M characters

Price listed in Microsoft AI's launch post of 1 October 2026.

Source: https://microsoft.ai/news/our-first-streaming-transcription-model/

## Strengths

- Generates 45 seconds of audio with 150ms end-to-end latency, per Microsoft's launch post
- 55% faster model inference, by Microsoft's account
- $15 per 1M characters, which Microsoft says is about 60% cheaper than comparable models
- Same 23 languages, 26 locales and cross-language voices as MAI-Voice-2.1
- Emotion and style control through SSML `mstts:express-as`, plus gated instant voice cloning from a 5–60 second reference clip
- Available in Azure Voice Live as well as the Azure Speech SDK, served from 14 Azure regions

## Best for

- Reach for it for real-time voice agents and assistants, the workload Microsoft built it for
- Reach for it for low-latency call-centre and IVR flows, a use Microsoft Learn lists for the model
- Reach for it as the speech output of a voice agent paired with MAI-Transcribe-2-Streaming, as Microsoft recommends
- Reach for MAI-Voice-2.1 instead for audiobooks and long-form narration, where fidelity matters more than speed

## How to access

- [Microsoft Foundry (Azure Speech)](https://learn.microsoft.com/en-us/azure/ai-services/speech-service/mai-voices) — model id `MAI-Voice-2.1-Flash`
- [OpenRouter](https://openrouter.ai/microsoft/mai-voice-2.1-flash) — model id `microsoft/mai-voice-2.1-flash`
- [Vercel AI Gateway](https://vercel.com/ai-gateway/models/mai-voice-2.1-flash) — model id `microsoft/mai-voice-2.1-flash`

## FAQ

### What is MAI-Voice-2.1-Flash?

It is the fast, low-latency variant of Microsoft AI's MAI-Voice-2.1 text-to-speech model, announced on 1 October 2026 for high-volume, latency-sensitive workloads such as voice agents and call-centre flows.

### How fast is MAI-Voice-2.1-Flash?

Microsoft says it can generate 45 seconds of audio with an end-to-end latency of 150ms, and that it delivers 55% faster model inference.

### What does MAI-Voice-2.1-Flash cost?

Microsoft's launch post lists $15 per 1M characters, which it describes as about 60% cheaper than comparable models. MAI-Voice-2.1 is listed at $22 per 1M characters.

### Does the Flash model support fewer languages than MAI-Voice-2.1?

No. Microsoft says it supports the same languages and cross-language speakers as MAI-Voice-2.1 — 23 languages and 26 locales — and every managed voice in the Microsoft Learn voice table supports both models.

### Can it clone a voice?

Yes, with gated access. Instant voice cloning uses a short consented reference clip (Microsoft Learn recommends 5–60 seconds); you apply through Microsoft's Limited Access Review, and only authorised, licensed voices can be synthesised.

### How do I use MAI-Voice-2.1-Flash?

On Microsoft Foundry it is in public preview: call it through the Azure Speech SDK or REST API with SSML and a voice name such as `en-US-Harper:MAI-Voice-2.1-Flash`, or through Azure Voice Live. Microsoft also lists MAI Playground, OpenRouter and Vercel AI Gateway (`microsoft/mai-voice-2.1-flash`), with LiveKit support coming soon.

## Links

- [Microsoft AI — Our first streaming transcription model debuts at no. 1 on Artificial Analysis (MAI-Voice-2.1-Flash announcement)](https://microsoft.ai/news/our-first-streaming-transcription-model/)
- [Microsoft Learn — MAI-Voice-2.1 and MAI-Voice-2.1-Flash](https://learn.microsoft.com/en-us/azure/ai-services/speech-service/mai-voices)
- [MAI Playground](https://playground.microsoft.ai/)
- [OpenRouter — MAI-Voice-2.1-Flash](https://openrouter.ai/microsoft/mai-voice-2.1-flash)
- [Vercel AI Gateway — MAI-Voice-2.1-Flash](https://vercel.com/ai-gateway/models/mai-voice-2.1-flash)

Source: AI/TLDR LLM registry, https://ai-tldr.dev/models/mai-voice-2-1-flash/
