Microsoft launched MAI-Transcribe-2-Streaming on October 1, alongside MAI-Voice-2.1 and its Flash variant. The company says transcription begins with partial hypotheses just over 100 milliseconds after receiving audio, then revises them as context arrives. Introductory transcription costs $0.54 per audio hour through year-end; the voice models cost $22 and $15 per million characters respectively. Early words can let an application begin work before a speaker finishes. They are still partial hypotheses, and Microsoft’s component latency claims do not measure the whole application.
Microsoft’s transcription now starts while you are speaking
Microsoft’s streaming transcription emits and revises hypotheses before speech ends.
1 source · 1 min read
Link to noteOriginally published 2 October 2026. Reporting, forecasts, prices, odds and event availability reflect the original edition.
Sources & further reading
- Read the original microsoft.ai