From creative exploration to enterprise workflows

Creating a compelling image is one step. Making image generation useful across a business means giving teams the control to refine an idea, adapt it for different audiences, and keep moving when the work scales. MAI-Image-2.6 and MAI-Image-2.6-Flash are now generally available in Microsoft Foundry, giving teams the precision and speed to match how creative work actually scales.

Since preview, we’ve been gathering customer feedback as teams explore where these models fit into their work. That dialogue is central to how we think about enterprise readiness: not just the first result, but consistency across edits, useful controls, and the right balance of quality and speed. We’re grateful to everyone who has shared examples, tested ideas, and helped us understand the demands of real creative workflows.

More context. More control.

The capabilities available across the family give teams more ways to guide an image from concept to usable asset. With multi-reference editing in MAI-Image-2.6, you can use up to five reference images to help maintain product, character, and brand-asset consistency through an edit. A creative team could bring together a product reference, a character, and a visual direction rather than describing every detail from scratch.

Web grounding adds relevant real-world context to generation through Bing integration, helping with prompts that call for recent or specific information. It’s another way to provide context, not a substitute for reviewing the output. For enterprise teams, that review remains important wherever an image needs to meet factual, brand, or editorial requirements.

Format and resolution controls, including dynamic aspect ratios and output up to 1.5K, support the practical work of adapting an idea for different placements. Together, these capabilities let teams explore how image generation can fit their creative process without treating every asset as an isolated request.

How are customers using the models

MAI-Image-2.6 is the choice for precision: complex edits, text rendering, and polished final campaign assets where the details matter. MAI-Image-2.6-Flash is suited to interactive creative tools, rapid iteration, personalization, and high-volume generation. Start with the requirements of your application, then evaluate which model best supports the experience you want to deliver.

A retail team generating seasonal campaign imagery might use this workflow: (1) upload three reference images, including a product photo, brand color palette, and target aesthetic; (2) generate 10 variations with different backgrounds using web grounding for trending seasonal themes; and (3) use the Flash model for initial exploration and revisions, then switch to MAI-Image-2.6 for final hero assets requiring text overlays.

Customers are already exploring how MAI-Image models can support these kinds of creative workflows. Minute, a leading global sports content and technology company, which operates the Sports Illustrated publishing business, has been testing the models across its creative use cases.

“At Minute, we're constantly exploring new technologies that can strengthen our creative capabilities across brands like Sports Illustrated. In our early testing of Microsoft's MAI Image models, we were particularly impressed by the visual quality, composition and precision of targeted design edits. We see exciting potential for these models to help our teams develop creative concepts, explore ideas and refine designs, while maintaining the creative control and standards our brands are known for,” said Noy Freedman, SVP of Strategic Alliances at Minute.

And soon, creators will be able to experience MAI-Image-2.6 in Adobe Firefly, bringing the model into the tools creators already use to move from ideas to polished assets.

“We’re excited to bring Microsoft AI’s Image 2.6 to Adobe Firefly soon, our all-in-one creative AI studio, to give creators even more choice, precision and control over how they bring their ideas to life,” said Matt Chotin, Senior Director of Product at Adobe.

We’re excited to see how customers and partners put these models to work, from focused creative applications to broader enterprise adoption.

Get started

Try it now: - Quickstart: Generate your first image (5 min) - Sample notebook: Multi-reference editing

Go deeper: - MAI-Image-2.6 model card - MAI-Image-2.6-Flash model card - Pricing calculator

MAI-Image-2.6 pricing starts at $5 USD per 1M tokens for text input, $8 USD per 1M tokens for image input, and $38 USD per 1M tokens for image output. MAI-Image-2.6-Flash pricing starts at $1.75 USD per 1M tokens for text input, $2.50 USD per 1M tokens for image input, and $19 USD per 1M tokens for image output.