Microsoft Slashes Speech AI Prices 72% With New MAI-Transcribe-2, Undercutting OpenAI and Google
Summary
Microsoft slashes speech AI prices 72% with its new MAI-Transcribe-2 model, offering transcription at just $0.10 per hour while undercutting OpenAI, Google, and ElevenLabs with support for 60 languages, advanced features, and top-tier accuracy — all part of a bold push to own its AI infrastructure and reduce reliance on outside partners.
Key Points
- Microsoft AI releases MAI-Transcribe-2, a speech-recognition model priced at $0.10 per hour — a 72% price cut from its predecessor launched just five months ago — undercutting OpenAI, Google, and ElevenLabs on both cost and speed.
- The model supports 60 languages, ranks second on the Artificial Analysis accuracy leaderboard, and bundles premium features like speaker diarization, word-level timestamps, keyword biasing, and code-switching for languages like Hinglish and Spanglish into its base price.
- Microsoft is aggressively building its own AI models to reduce dependence on OpenAI, quietly replacing third-party models in products like Word, Excel, and Teams, as part of a broader strategy to own AI infrastructure across every modality rather than paying external partners.