Microsoft introduces MAI-Transcribe-2 transcription model

Microsoft introduces MAI-Transcribe-2, a transcription model that the company says is the fastest, cheapest and most accurate in the world.

Mustafa Suleyman, Microsoft's post on X, says the model runs 10 times faster than GPT-Transcribe and 5 times faster than Gemini 3.5, and costs $0.10 per hour of audio. These speed and price comparisons are Microsoft's own claims and have not been independently verified in the report.