Microsoft takes on AI rivals with three new foundational models
Why it matters: Microsoft is directly challenging Google and OpenAI in the LLM market with new, cheaper models, potentially shifting market share.
- Microsoft AI released three foundational AI models: MAI-Transcribe-1 (speech-to-text in 25 languages, 2.5x faster than Azure Fast), MAI-Voice-1 (audio generation, 60 seconds in one second), and MAI-Image-2 (video generation).
- Mustafa Suleyman, CEO of Microsoft AI, stated the company's 'Humanist AI' approach prioritizes human-centered design and practical use, with more models expected soon in Foundry and Microsoft products.
- Microsoft's MAI Superintelligence team developed these models, which aim to be cheaper than offerings from Google and OpenAI, with MAI-Transcribe-1 starting at $0.36/hour and MAI-Voice-1 at $22/million characters.
- Suleyman reaffirmed Microsoft's commitment to its partnership with OpenAI, despite launching its own models, noting to The Verge that a recent renegotiation of the partnership enabled this superintelligence research push.
Microsoft AI has launched three new foundational AI models—MAI-Transcribe-1, MAI-Voice-1, and MAI-Image-2—to generate text, voice, and images, intensifying its competition with rivals like Google and OpenAI while still maintaining its partnership with OpenAI, a relationship recently renegotiated to allow for this superintelligence research. These models, developed by the MAI Superintelligence team led by Mustafa Suleyman, emphasize a 'Humanist AI' approach, focusing on practical, human-centric applications and offering more affordable pricing than competitors.
