Microsoft Debuts Its First In-House AI Image Generator, MAI-Image-1, on Bing and Copilot

Microsoft has officially launched MAI-Image-1, its first in-house AI image generator, marking a significant milestone in the company’s growing AI portfolio. The model is now integrated into Bing Image Creator and Copilot Audio Expressions, though it is not yet accessible in the European Union.

The new model, introduced in October, aims to deliver high-quality, photorealistic visuals with remarkable efficiency. Microsoft AI chief Mustafa Suleyman announced the rollout on X (formerly Twitter), describing MAI-Image-1 as a breakthrough in AI-generated imagery. He noted that the model “really excels at” creating images of food, nature scenes, artsy lighting, and photorealistic detail, underscoring its creative range and visual finesse.

According to Microsoft’s official blog, MAI-Image-1 demonstrates exceptional performance in generating lighting effects such as bounce light and reflections, as well as realistic landscapes. The company claims the model’s balance of speed and quality enables users to bring their ideas to life faster, iterate quickly, and refine their visuals using other creative tools.

“MAI-Image-1 excels at generating photorealistic imagery, like lighting (e.g., bounce light, reflections), landscapes, and much more. This is particularly so when compared to many larger, slower models. Its combination of speed and quality means users can get their ideas on screen faster, iterate through them quickly, and then transfer their work to other tools to continue refining,” Microsoft stated.

Beyond static image generation, the new AI model will also be featured in Copilot’s text-to-speech platform, Copilot Audio Expressions, where it will create complementary visuals for AI-generated audio stories in “story mode.” This integration reflects Microsoft’s broader ambition to merge different AI modalities — from text and voice to visuals — into cohesive creative experiences.

The debut of MAI-Image-1 follows Microsoft’s earlier introduction of its first in-house models in August: the speech model MAI-Voice-1 and the text-based MAI-1-preview. At that time, Microsoft hinted that MAI-1-preview would be incorporated into Copilot, suggesting a gradual shift away from total reliance on OpenAI’s technologies.

As of now, Microsoft confirms that Copilot continues to operate using OpenAI’s latest model, GPT-5, while also providing access to Anthropic’s Claude models for users seeking alternatives. This diversified AI ecosystem indicates Microsoft’s strategy to combine internal innovation with partnerships across the broader AI landscape.

Currently, MAI-Image-1 appears as one of three available AI models on the Bing Image Creator platform and app. The other two are OpenAI’s DALL-E 3 and GPT-4o, offering users a variety of creative engines to explore.

By developing MAI-Image-1, Microsoft is positioning itself as a serious contender in the generative AI space, reducing dependence on external providers while enhancing its ecosystem with faster, more efficient creative tools. The move highlights the company’s ongoing investment in homegrown AI research — and signals that the next phase of AI-powered creativity may increasingly come from within Redmond’s own labs.