MAI-Image-2.5: Generate and edit images with precise scene control

mai-image-2.5 Microsoft AI’s new image generation and editing model is now ranked No. 2 on Arena’s image-edit leaderboard and No. 3 for text-to-image as of June 1, 2026.

Most image models treat editing as regeneration, swapping out one thing, risking breaking everything else. MAI-Image-2.5 sees it differently: localized edits that understand the visual context, so changing the background, replacing text, or adding objects doesn’t spoil what you didn’t touch.

🖼 text-to-image generation With strong prompt following and text rendering so that your prompts produce exactly what you described, not an interpretation of it

✂️ localized editing in objects, backgrounds and text without affecting the surrounding image

🧑 face and identity uniformity Preserve across pose, expression and viewpoint changes useful for product images and business workflows

mai-image-2.5-flash Nearly half the output token cost for high-throughput, cost-sensitive pipelines

Built for developers and ML teams, embedding image creation or editing into production apps where controllability, identity management, and cost-performance tradeoffs all matter.

Try it in MAI Playground or access it through Azure Foundry and OpenRouter today.

PS I’m looking for the latest and greatest launches in tech, SaaS and AI, follow to stay informed @rohan recommends



<a href

Leave a Comment