Microsoft’s New MAI-Image-2.6-Flash Is More Than Twice as Fast and Far Cheaper

Microsoft MAI-Image-2.6 is now available to developers through Microsoft Foundry, alongside a new Flash model designed for faster and cheaper image generation.

Both models support image generation and editing, reference images, flexible output formats, and production-focused workloads.

Microsoft brings MAI-Image-2.6 to Foundry

Microsoft has made MAI-Image-2.6 available to developers through Microsoft Foundry.

The model focuses on high-quality image generation, complex editing, improved text rendering, and greater control over the final output.

MAI-Image-2.6-Flash focuses on speed and lower costs

Alongside the standard model, Microsoft has introduced MAI-Image-2.6-Flash, which targets applications that need faster responses and higher throughput.

Microsoft says Flash generates images more than twice as fast as GPT-Image-2-Medium while delivering 72% greater efficiency.

The company recommends the model for a range of workloads, including interactive applications, rapid creative iteration, personalized image generation, automated content pipelines, and high-volume production tasks.

Flash keeps the core MAI-Image-2.6 capabilities

MAI-Image-2.6-Flash supports text-to-image generation as well as image-to-image editing.

Developers can make targeted changes to individual objects, text, layouts, or other parts of an image without recreating the entire visual.

Both models can also accept up to five reference images in a single request. This allows developers to maintain consistency across products, characters, people, branding, and other recurring visual elements.

The models also support dynamic aspect ratios and higher-resolution output, giving developers more control over images designed for different formats and platforms.

MAI-Image-2.6 can use Bing Search for web grounding

Microsoft has also added web grounding to MAI-Image-2.6, allowing the model to retrieve current real-world information through Bing Search.

This can improve image generation when prompts depend on current information or accurate details about real-world places, objects, and subjects instead of relying entirely on the model’s training data.

MAI-Image-2.6-Flash costs significantly less

Microsoft has priced MAI-Image-2.6-Flash at $1.75 per million text input tokens, $2.50 per million image input tokens, and $19 per million image output tokens.

That makes Flash roughly three times cheaper for text and image input, while image output costs half as much compared to MAI-Image-2.6.

Microsoft recommends Flash when developers prioritize latency, throughput, and cost. The standard MAI-Image-2.6 remains the better choice for complex edits, commercial designs, final production assets, and workloads that require the highest image quality and accurate text rendering.

Both models are now available in public preview

MAI-Image-2.6 and MAI-Image-2.6-Flash are now available in public preview through Microsoft Foundry, giving developers access to Microsoft’s latest in-house image generation technology.

The release also expands Microsoft’s growing MAI model lineup. The company recently launched MAI-Transcribe-2, its latest speech recognition model, with a focus on faster and cheaper transcription.