Microsoft launches MAI-Transcribe-2-Streaming real-time transcription AI for $0.54 per hour
Microsoft’s AI division announced MAI-Transcribe-2-Streaming, a real‑time transcription model that streams provisional text and finalizes it after each sentence. It supports 60 languages, auto‑detects the spoken language, and offers speaker diarization, word‑level timestamps, and custom term recognition. Microsoft says the model ranked first out of 38 on the Artificial Analysis evaluation site…
What you can do with it
AI at Work →Generate meeting subtitles in real time as people speak
What you get You can follow meeting discussions and create live notes.
Converts incoming audio into text sequentially with intermediate results that update.
- Who for
- marketers, sales, founders and support teams
- Cost
- paid from $0.54
- Effort
- minutes
Use it for
- Follow meeting discussions
- Create live notes
- Check speaker clarity
How to set it up · From the article
- Set up Azure contract and configuration for supported regions
- Use short audio clips to test accuracy and timing
- Verify language support and consent requirements before full use
Watch out Limited rollout: not open to everyone yet; Not available in Japan
Not for everyone not in Japan
MAI-Transcribe-2-Streaming in the tool directory · 2 changes covered
Key points from the news
- Microsoft released MAI-Transcribe-2-Streaming, a real‑time transcription model supporting 60 languages, priced at $0.54 per hour (introductory price).
- In Artificial Analysis evaluation (Sept 28) it ranked 1st of 38 models with 2.5% error rate and 0.13 s latency to finalization.
- The model is in public preview; Microsoft advises against production use and highlights possible errors in proper nouns, numbers, and unverified language support such as Japanese.
The service is available in public preview through Microsoft Foundry, and Microsoft advises against production use. Errors, especially in proper nouns, numbers, and dates, can have serious consequences for official records, and accuracy may vary by language, accent, and background noise. Japanese support is listed among the 60 languages but has not been independently verified. Companies are urged to pilot the tool in a single scenario, have humans verify critical details, and establish clear recording notifications and retention policies.
The article notes that while the low cost removes a financial barrier, the preview status and potential transcription mistakes mean organizations must design verification workflows before relying on the AI for official documentation.
The story so far
2 episodes →- Microsoft launches MAI-Transcribe-2-Streaming real-time transcription AI for $0.54 per hourthis story
Turning speech into text as it happens for $0.54 per hour — Microsoft's new transcription AI and what to decide before entrusting it with meetings and calls
note.com · 4 October 2026
Loading the full article…
This text was published by note.com and written by たいち. It is reproduced here with attribution so you can read it in full; the rights remain with the publisher. Read it at the source ↗
Coverage and discussion
3sourcesThe headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.
More in Generative AI & Models
All →- Apiiro finds 1,122 public GitHub files with secrets saved by Claude Code · 1 src
- Thinking Machines releases Inkling 2 Small, a 594B MoE model with 17B active parameters · 8 src
- OpenAI releases GPT-6 Cyber, a Critical-rated model that outperforms predecessors in vulnerability detection and alignment · 25 src
- Google releases Gemini Omni 1.2 Flash with 15-second video generation and 30% price cut to $0.07 per second · 10 src
- US accuses Chinese executive of allegedly sharing Nvidia chip info with Beijing · 11 src
Comments
via GitHub Discussions