tech
April 3, 2026
Microsoft launches three in-house AI models in direct challenge to OpenAI
Six months after renegotiating the contract that once barred it from independently pursuing frontier AI, Microsoft has released three in-house models that directly challenge the partner it spent $13 billion cultivating. MAI-Transcribe-1, MAI-Voice-1, and MAI-Image-2 are now available in Microsoft Foundry, and they do not carry OpenAI’s name anywhere on the label.
TL;DR
- Microsoft launched MAI-Transcribe-1, MAI-Voice-1, and MAI-Image-2, developed by its MAI Superintelligence team.
- These models compete directly with OpenAI's offerings, particularly after a contract renegotiation allowed Microsoft to pursue independent AI development.
- MAI-Transcribe-1 claims superior performance in speech-to-text across multiple languages compared to OpenAI's Whisper and Google's Gemini.
- MAI-Voice-1 generates natural-sounding speech and supports custom voice creation, enabling a complete voice pipeline without OpenAI.
- MAI-Image-2 ranks highly on the Arena.ai text-to-image leaderboard.
- The development reflects a broader fragmentation in the AI market, with companies like Anthropic and Google also rapidly iterating.
- Microsoft's Foundry platform hosts both its own and OpenAI models, but the new releases aim to shift enterprise spending towards in-house solutions.
- While the partnership with OpenAI continues, Microsoft is strategically dismantling the premise that it needs OpenAI to compete in AI.
Continue reading the original article