Microsoft introduces its own image generator MAI-Image-1

Microsoft AI has announced the launch of MAI-Image-1, its first internally developed text-to-image generation model. The company calls it the next step in building its own AI ecosystem, following MAI-Voice-1 and MAI-1-preview, aimed at reducing dependence on OpenAI.

During development, Microsoft collaborated with designers and photographers to avoid repetitive or overly generic results. According to the company, MAI-Image-1 delivers some of the most realistic images on the market—especially in natural scenes, lighting, and landscapes—and operates faster than larger generative systems, making it suitable for high-demand products.

The model has already entered the top 10 on LMArena, a platform where users compare and vote on AI-generated images. Microsoft plans to integrate MAI-Image-1 into Copilot and Bing Image Creator soon.

The company emphasizes its commitment to safety and responsible AI use, working on content verification systems to ensure ethical operation.