Microsoft AI released MAI-Transcribe-2, a speech recognition model supporting 60 languages. The system includes speaker diarization to identify individual voices in recordings.
Microsoft claims the tool is five times faster than Google’s Gemini 3.5 Transcribe. It reportedly operates ten times faster than OpenAI’s GPT-Transcribe.
Promotional pricing is set at $0.10 per audio hour through 2026. This rate represents a 72% price reduction from Microsoft’s previous version.
The strategy aims to capture enterprise market share from Google and OpenAI. This launch signals Microsoft’s shift toward proprietary models to reduce partner reliance.