MAI Image 2.7
Image Generator & Editor
An idea is all it takes. Create from a prompt, or upload your images and tell MAI what to change. Two models. One creative space.
Credits reserved upfront. Final cost follows usage, capped at this quote; unused credits are returned.
New accounts receive 10 welcome credits. Each generation requires enough credits for its displayed reservation.
A little inspiration. A lot of possibility.
Explore text, portraits, products and places. Choose a prompt to make it your own.
Examples published by OpenRouter for MAI Image 2.7 / Flash. Imported references, not ImageLayered test results. OpenRouter ↗
From first thought to final image
Start with words. Bring a reference. Make it yours.
Describe or upload
Write your idea, or drop in up to five images to switch straight into editing.
Choose your MAI
Use the precision model for controlled detail, or Flash to explore ideas at a lower cost.
Create, refine, download
Generate an image, download the original, or use it as the reference for your next edit.
One studio. More room to create.
A prompt is your starting point
Describe a product scene, portrait, illustration or visual concept in natural language.
Bring your references together
Combine up to five images for people, products, styles and scenes in one edit.
Change what matters
Ask for a different setting, palette or material, and describe the details you want preserved.
Choose your composition
Pick square, portrait or landscape from seven fixed ratios, or leave the composition to Auto.
Precision or Flash
Choose the precision tier for detailed work, or the lower-cost Flash variant for faster iteration.
Your results, ready to return to
Completed images appear in Recent Generations for 30 days, with private preview and download.
Made for the way you create
Product & brand visuals
Build a new setting around a product reference, then refine the color and lighting.
Portraits & personal projects
Explore portrait direction, clothing ideas and illustrative treatments with your own photos.
Social & campaign concepts
Try different compositions and visual directions before choosing a finished campaign image.
Design & visual storytelling
Combine subjects, scenes and style references into a cohesive visual idea.
Meet MAI Image 2.7. Then take your idea further.
MAI Image 2.7 is Microsoft AI’s generation and editing family. The precision tier focuses on controlled, iterative work; Flash offers a lower-cost, lower-latency option. Both accept text and up to five image references. Choose a fixed ratio or let the model compose automatically.
Give clear instructions about what to change and what to keep. Reference images guide the result; exact identity, lettering and fine details still need your review. ImageLayered brings both MAI models into the same workspace as our other image and video tools.
Read the model overviewExplore more image models
Give your next idea motion
MAI Image 2.7, explained
A few things to know before you create.
What is MAI Image 2.7?
MAI Image 2.7 is Microsoft AI’s image generation and editing model. On ImageLayered you can select the precision model or MAI Image 2.7 Flash, write a prompt, and optionally supply up to five references.
How do I switch from generation to editing?
Upload, drop or paste an image into the hero composer. It switches to Edit Image automatically. Add up to five references and describe the change. Remove all references to return to text-to-image.
How is Flash different from the precision model?
Flash is positioned as the lower-cost, lower-latency variant. Both models accept text and images, with up to five references and the same aspect ratio choices. Results depend on your prompt and references.
Is MAI Image 2.7 free?
New ImageLayered accounts receive 10 welcome credits. Each request needs enough balance for the reservation shown before generation. Charges follow actual usage and unused reserved credits are returned. Additional credits are available through paid plans or packs; this is not unlimited free generation.
Which settings can I use?
Choose Auto or one of seven fixed aspect ratios, plus a solid or transparent background. Each request returns one image. This integration does not offer separate quality or resolution tiers.
What happens if I close the page or generation fails?
Accepted requests continue in the background. Reopen this page to restore your latest task, or find saved results in Recent Generations for 30 days. Failed or timed-out tasks receive an automatic refund of their reserved credits.
Can I use multiple references?
Yes. Attach up to five JPG, PNG, WebP or AVIF files, each up to 10 MB. Refer to them as Image 1, Image 2 and so on in your prompt. Only upload images you are entitled to use.
Your next image starts with an idea.
Write it. Upload it. Make it yours with MAI.
Start creating