Microsoft dropped three in-house AI models on Thursday — MAI-Transcribe-1, MAI-Voice-1, and MAI-Image-2 — and the timing is hard to miss. The company still holds an OpenAI stake worth roughly $135 billion, yet it’s now shipping direct competitors to OpenAI’s own transcription, voice, and image products. MAI-Transcribe-1 clocks a 3.8% average Word Error Rate across 25 languages on the FLEURS benchmark, beating OpenAI’s Whisper-large-v3 on all 25. Meanwhile, MAI-Voice-1 can produce a full minute of synthetic speech in under a second, and MAI-Image-2 runs at least twice the generation speed of its predecessor. All three are live on Microsoft Foundry right now.
(Also Read : NASA’s Artemis II Launch – lunar mission in 54 years)
What makes this interesting isn’t the specs — it’s the positioning. Microsoft spent years packaging OpenAI’s models inside Copilot and Azure; now it’s building and shipping more of its own foundational stack. MAI-Transcribe-1 handles noisy and low-quality audio across multiple languages, making it useful for meeting transcription and call-center analytics. MAI-Image-2 is already rolling out in Bing and PowerPoint. The pricing is competitive too — MAI-Transcribe-1 starts at $0.36 per hour, MAI-Voice-1 at $22 per million characters. Whether these models are good enough to pull serious workloads away from Google or OpenAI is still an open question, but Microsoft clearly isn’t waiting around to find out.
For Latest Tech Updates Please follow us :
facebook | twitter | instagram | telegram | whatsapp