How much does MAI-Image-2.6 cost?

MAI-Image-2.6 is billed per token on OpenRouter through the Azure provider at $5.00 per million text input tokens, $8.00 per million image input tokens and $38.00 per million image output tokens. Microsoft's own model page does not print rates and points to Microsoft Foundry for production pricing. Microsoft positions the Flash variant as production-ready quality at less than half the price.

What is the difference between MAI-Image-2.6 and MAI-Image-2.6-Flash?

MAI-Image-2.6-Flash is the latency-optimised sibling aimed at high-throughput production workloads. Microsoft measures Flash generating images 2.8x faster than GPT-Image-2-Medium with 72% greater efficiency, and prices it at less than half the full model. MAI-Image-2.6 stays the stronger option when final image quality matters more than speed or cost.

How does MAI-Image-2.6 compare to MAI-Image-2.5?

Microsoft reports MAI-Image-2.6 improving +79 Elo over MAI-Image-2.5 overall on Arena, with gains in every text-to-image category. Text rendering improves the most, at +91 Elo. Microsoft also cites better portraits, 3D images and reference-image support, and says the result puts the MAI-Image family ahead of leading models from Meta, Google and xAI.

Where can I use MAI-Image-2.6 today?

MAI-Image-2.6 is in public preview on Microsoft Foundry and can be tried in MAI Playground. Both the standard and Flash variants were listed on OpenRouter on 4 September 2026 through the Azure provider, so an application that already routes requests through OpenRouter can call the model without setting up a separate Foundry account first.

Microsoft AI · 2026-09-04 · major

MAI-Image-2.6 — Microsoft's image model lands at No. 2 on Arena

MAI-Image-2.6 is Microsoft AI's new image generation and editing model, now second on the Arena text-to-image and image-editing boards. A Flash variant makes pictures 2.8x faster than GPT-Image-2-Medium.

Microsoft AI announcement card for MAI-Image-2.6

Microsoft's newest image model ranks second on Arena and ships with a faster, cheaper Flash variant.

Key specs

Elo gain over mai image 2.5	+79
Flash speed	2.8x faster than GPT-Image-2-Medium

Quick facts

Maker	Microsoft AI
Variants	MAI-Image-2.6 and MAI-Image-2.6-Flash
Arena rank	No. 2 for text-to-image and image editing
Artificial Analysis	No. 1 image editing, No. 2 text-to-image
Max resolution	Up to 1.5K
Availability	Public preview in Microsoft Foundry; MAI Playground
Previous version	MAI-Image-2.5

Pricing

Text input	$5.00 / 1M tokens
Image input	$8.00 / 1M tokens
Image output	$38.00 / 1M tokens

source ↗

What is it?

MAI-Image-2.6 adds multi-image reference editing, web grounding and dynamic aspect ratios to Microsoft AI's image line, at resolutions up to 1.5K. A second variant, MAI-Image-2.6-Flash, is built for latency-sensitive, high-throughput production work. Both went into public preview on Microsoft Foundry and are also in MAI Playground.

How does it work?

Web grounding lets the model pull current web context into a picture instead of relying only on what it learned in training. Multi-image reference support means you can hand the model several pictures at once and keep a style or subject consistent across the result. Microsoft describes MAI-Image-2.6 as its strongest image model and Flash as the same family tuned for throughput.

Why does it matter?

Arena and Artificial Analysis both place the model second for text-to-image, and Artificial Analysis puts it first for image editing — so a team picking an image API now has a Microsoft option at the top of the table. Microsoft reports a +79 Elo gain over MAI-Image-2.5 overall, with text rendering alone up +91 Elo, which is the weak spot most image models still have.

Who is it for?

design and marketing teams, image API developers

Frequently asked questions

How much does MAI-Image-2.6 cost?: MAI-Image-2.6 is billed per token on OpenRouter through the Azure provider at $5.00 per million text input tokens, $8.00 per million image input tokens and $38.00 per million image output tokens. Microsoft's own model page does not print rates and points to Microsoft Foundry for production pricing. Microsoft positions the Flash variant as production-ready quality at less than half the price.
What is the difference between MAI-Image-2.6 and MAI-Image-2.6-Flash?: MAI-Image-2.6-Flash is the latency-optimised sibling aimed at high-throughput production workloads. Microsoft measures Flash generating images 2.8x faster than GPT-Image-2-Medium with 72% greater efficiency, and prices it at less than half the full model. MAI-Image-2.6 stays the stronger option when final image quality matters more than speed or cost.
How does MAI-Image-2.6 compare to MAI-Image-2.5?: Microsoft reports MAI-Image-2.6 improving +79 Elo over MAI-Image-2.5 overall on Arena, with gains in every text-to-image category. Text rendering improves the most, at +91 Elo. Microsoft also cites better portraits, 3D images and reference-image support, and says the result puts the MAI-Image family ahead of leading models from Meta, Google and xAI.
Where can I use MAI-Image-2.6 today?: MAI-Image-2.6 is in public preview on Microsoft Foundry and can be tried in MAI Playground. Both the standard and Flash variants were listed on OpenRouter on 4 September 2026 through the Azure provider, so an application that already routes requests through OpenRouter can call the model without setting up a separate Foundry account first.

Try it

https://playground.microsoft.ai/chat?model=mai-image-2-6