Skip to content
Türkçe

Microsoft · Models · Agents

Microsoft releases three new voice models, one transcribing 60 languages in real time

Published: 5 sourcesTürkçe

Microsoft AI announced three models for voice agents on October 1. MAI-Transcribe-2-Streaming, the company’s first streaming transcription model, turns speech into text in real time across 60 languages with automatic language detection, and Microsoft says it ranks first for accuracy on Artificial Analysis. The text-to-speech models MAI-Voice-2.1 and the faster MAI-Voice-2.1-Flash cover 23 languages, with Flash quoted at 150 ms end-to-end latency. Both can clone a voice from a few seconds of audio and include consent safeguards. The models are offered through Microsoft Foundry, with transcription priced at $0.54 per audio hour until year end.

Sources

  1. Microsoft AI, “Our first streaming transcription model debuts at no. 1 on Artificial Analysis”, (microsoft.ai)
  2. Microsoft AI, “MAI-Transcribe-2-Streaming Model Card”, (microsoft.ai)
  3. Microsoft AI, “MAI-Voice-2.1-Flash Model Card”, (microsoft.ai)
  4. Microsoft AI (X), Microsoft AI post on X (announcing the three models), (x.com)
  5. Artificial Analysis (X), Artificial Analysis post on X (AA-WER Streaming ranking), (x.com)

About this story

This story was posted on Instagram by @jarrus.tech on Oct. 10, 2026.

Spotted an error in this story? [email protected] · Instagram

This story in Turkish: Microsoft, biri 60 dilde gerçek zamanlı yazıya döken üç yeni ses modeli çıkardı

On the same topic