Microsoft launches MAI-Transcribe-2: single multilingual transcription model with diarisation and timestamps
mustafasuleyman · x · 2026-09-12
Microsoft AI chief Mustafa Suleyman announced MAI-Transcribe-2, a single model for highly accurate multilingual audio transcription across accents and challenging audio conditions, with built-in speaker diarisation and word-level timestamps.
It supports keyword biasing for domain-specific terminology in fields like pharma and medicine, targeting captions, call analysis, and accessibility workflows out of the box. A playground demo is live.
More from Models
- Kording Lab: AI credits inventors but almost never credits the original source of an idea — KordingLab · 2026-09-12
- DeepSeek Flash 4.1 Draws Early Praise From Developer Who Tried It — _arohan_ · 2026-09-12
- The real worry in Anthropic's report: no way to stop distillation by rivals — kimmonismus · 2026-09-12
- Meta's Confidential Computing Promise Doesn't Protect Data if Inference Isn't Attested — signulll · 2026-09-12
- Among a Flood of New AI Labs, the Humans Team Ships Its First Release — niloofar_mire · 2026-09-12
- Agent traces reveal hour-long costly 'super resolution' rabbit hole in continual learning study — yuxiangw_cs · 2026-09-12