Microsoft Launches Its First In-House AI Image Generator, MAI-Image-1, Now Available in Bing and Copilot for Creative Users and Developers

Microsoft’s first in-house AI image generator, MAI-Image-1, is out now. You can try it in Bing Image Creator and Copilot Audio Expressions. It’s also available in all countries that support these services, except the EU for now.

On Bing, you can find it at bing.com/create, on the mobile app, or from the search bar. It’s one of the options in the model menu, next to DALL-E 3 and GPT-4o (both from OpenAI). Meanwhile, on Copilot, it’s now part of Story Mode. MAI-Image-1 will visualize your stories in Audio Expressions.

It was announced on October 13, when it debuted among the top 10 text-to-image models on LMArena. LMArena is an AI benchmark platform where users compare and rate large language models (LLMs) based on their outputs. MAI-Image-1 is currently number nine on the board, tied with Seedream 3.

Microsoft AI says the developers “prioritized 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.”

You can find several examples on Microsoft AI’s blog.

Overall, it’s a generator that puts creators’ needs first. It avoids the typical pitfalls of low-quality models, such as generic or repetitive designs.

Specifically, MAI-Image-1’s area of expertise is photorealistic outputs. It quickly delivers resource-intensive designs such as detailed landscapes and realistic lighting, with reflections and bounce lights. Most other AI tools either struggle with these tasks or are relatively slow.

Microsoft Is Building Its Own AI Models Now

Interestingly, this debut is yet another sign of Microsoft gradually drifting apart from OpenAI. The company played a big role in helping OpenAI build a name for itself in the industry. But now, it’s directing more and more of its AI efforts in-house, while OpenAI is collaborating more with other partners.

In August, Microsoft AI announced the first two in-house AI models: MAI-Voice-1 and MAI-1-preview. MAI-Voice-1 is a natural speech generation model, available on Copilot Labs, Daily, and Podcasts.

Furthermore, MAI-1-preview is a chatbot model, initially intended to roll out as part of Copilot. However, Copilot Chat is currently transitioning to OpenAI’s GPT-5 as its primary supporting LLM. It’s also adding Anthropic’s Claude AI as an optional supporting LLM.

Microsoft’s souring relationship with OpenAI aside, MAI-Image-1 is an exciting addition to its AI arsenal. If you’re a creator who needs a steady supply of photorealistic designs, it might be worth your time. Feel free to check it out now on Bing and Copilot.