HaiAI123

Curated Global AI Tools Directory

Industry update
25
heat index
Flat

Microsoft AI releases MAI-Transcribe-2-Streaming real-time transcription model

10/02/2026 — 10/03, 14:12·4 sources (all)·4 reports (all)·In the 2026-10-03 briefing

Key takeaways

  • Microsoft has added MAI-Transcribe-2-Streaming, which supports real-time transcription in 60 languages, alongside two multilingual speech mo…

Background & analysis

On October 2, 2026, Microsoft expanded its MAI model family with its first streaming transcription model, debuting alongside two other models focused on text-to-speech. According to SiliconANGLE, the release is aimed at developers who want to build voice agents that can listen to people's voices and reply instantly, similar to how humans talk to one another. The same day, The Decoder identified the transcription model as MAI-Transcribe-2-Streaming and noted that the batch targets the transcription and synthesis stages that voice agents require.

Later that day, TestingCatalog added further detail: MAI-Transcribe-2-Streaming handles real-time transcription across 60 languages, the two accompanying voice models are multilingual, and streaming audio is priced at $0.54 per hour.

On October 3, MarkTechPost reported that MAI-Transcribe-2-Streaming is the first real-time speech-to-text model released by Microsoft AI, and that it ranks #1 of 38 models on Artificial Analysis's AA-WER Streaming leaderboard. The model scores 2.5% WER at 0.13s on final transcripts and 2.5% at 0.12s on first partials. It covers 60 languages with continuous language detection and is available now in public preview in Microsoft Foundry. On pricing, the two later reports diverge slightly in wording: TestingCatalog states the $0.54 per hour figure for streaming audio directly, while MarkTechPost frames that same $0.54 per hour as an introductory-period price. Across the coverage, Microsoft AI is consistently named as the releasing organization, and the model's name, language coverage and availability are reported the same way.

AI-generated from 4 reports · updated 50 days ago

Latest turnMicrosoft AI has released MAI-Transcribe-2-Streaming, its first real-time speech-to-text model, which tops Artificial Analysis' AA-WER Streaming leaderboard of 38 models. It posts 2.5% WER at 0.13s on final transcripts and 2.5% at 0.12s on first partials, covers 60 languages with continuous language detection, and costs $0.54 per hour during the introductory period; it is in public preview on Microsoft Foundry.

updream
24-hour heatHeat index 25
Just charted — trend is building up
24 hours agonow

Reports on this story headlines open the original

Oct 3
  1. Microsoft AI has released MAI-Transcribe-2-Streaming, its first real-time speech-to-text model, which tops Artificial Analysis' AA-WER Streaming leaderboard of 38 models. It posts 2.5% WER at 0.13s on final transcripts and 2.5% at 0.12s on first partials, covers 60 languages with continuous language detection, and costs $0.54 per hour during the introductory period; it is in public preview on Microsoft Foundry.

    MarkTechPostAI score 85
Oct 2
  1. Microsoft has launched MAI-Transcribe-2-Streaming, bringing real-time transcription in 60 languages, along with two multilingual voice models. Streaming audio is priced at $0.54 per hour.

    TestingCatalogAI score 68
  2. Microsoft AI has released MAI-Transcribe-2-Streaming, a new model for real-time transcription, along with text-to-speech models built for voice agents. Both target the transcription and speech-synthesis layers that voice agents rely on, though Microsoft has not detailed availability or benchmark results.

    The DecoderAI score 70
  3. Microsoft expanded its MAI model family with its first streaming transcription model, launched alongside two text-to-speech models. They target developers building voice agents that can listen and reply instantly, the way people talk to each other.

    SiliconANGLEAI score 76

Other stories people are talking about

How is heat calculated?About the method

Heat counts how many independent sources covered a story in the last 48 hours: one source counts once no matter how many posts it published, decaying with a 24-hour half-life. What ranks first is what many people are talking about.

This page aggregates public feeds. Headlines and summaries are machine-organized and remain the property of the original authors; verify important facts at the source. If you believe a headline or summary infringes your rights, email the contact address in the footer with the page URL and basis of your claim, and we will remove or replace it after review.

Surge
Discussion rising fast
New
First report within 6 hours
Rising
Still gathering discussion

Back to AI industry updates →