Microsoft has announced its first text-to-image generator, MAI-Image-1, designed and developed in-house. Though Microsoft works closely with OpenAI, this is an important move from the Satya Nadella-led company, which announced its first two in-house models in August.
“We trained this model with the goal of delivering genuine value for creators, and we put a lot of care into avoiding repetitive or generically stylised outputs. For example, we prioritised rigorous data selection and nuanced evaluation focused on tasks that closely mirror real-world creative use cases — taking into account feedback from professionals in the creative industries. This model is designed to deliver real flexibility, visual diversity, and practical value,” the company said.
Microsoft says it sought feedback from creative professionals to avoid “repetitive or generically stylised outputs.” MAI-Image-1 “excels” at photorealistic imagery like lightning, landscapes, and more. It can process requests and produce images faster than “larger, slower models”. The model has already secured a spot in the top 10 of LMArena.
MAI-Image-1 joins Microsoft’s other AI products — voice generator MAI-Voice-1 AI and chatbot MAI-1-preview. The company has recently started using Anthropic’s AI models for some features in Microsoft 365 and is making “significant investments” in training its own AI models like MAI-Image-1. MAI-Image-1 will be available in Copilot and Bing Image Creator very soon.