Microsoft has launched MAI Image 2.7 and MAI Image 2.7 Flash, two new AI image generation and editing models designed to deliver high-quality visuals with improved precision, consistency and lower costs. The flagship model is focused on advanced image generation and detailed editing, while Flash is designed for faster image creation and production-scale workloads.

The launch is part of Microsoft’s broader expansion of its in-house MAI model family, which now spans coding, image generation, cybersecurity, speech and transcription. Microsoft is positioning MAI Image 2.7 as a competitive alternative to leading image models, claiming it ranks first in image editing evaluations while costing approximately one-fifth as much as a leading competitor.

MAI Image 2.7 Ranks No. 1 in AI Image Editing

Microsoft says MAI Image 2.7 secured the top position on Artificial Analysis’ image-editing leaderboard at launch, outperforming GPT Image 2.6 Sunburst in the evaluated image-editing category.

The model is designed to perform targeted edits without unnecessarily changing other parts of an image. Users can modify objects, alter materials, change colours, reposition elements, adjust lighting and refine backgrounds while retaining the original visual composition.

This capability addresses one of the major challenges in generative AI image editing: maintaining consistency when a user makes multiple changes to an existing image.

Instead of regenerating an entire image after every adjustment, users can make more precise modifications to individual elements.

Microsoft says the model also performs strongly in image-generation evaluations, placing among the leading models across image generation and editing leaderboards.

The company is positioning MAI Image 2.7 as a model suitable for creative workflows that require greater control over the final visual output.

Targeted Editing Without Recreating the Entire Image

One of the central features of MAI Image 2.7 is its ability to make localised changes to an image.

Users can ask the model to change the colour of an object, remove an unwanted element, modify the background, reposition a subject or improve a specific area while preserving the rest of the composition.

For example, a marketer could change the packaging colour in a product advertisement without recreating the entire scene. A designer could adjust lighting, replace a background object or modify the material of a product while keeping the overall image consistent.

The model is also designed to handle artifact cleanup, including unwanted visual distortions and motion-blur-related issues.

Such editing capabilities could reduce the number of repeated generations required to achieve a desired result, making AI image workflows more efficient for creators and businesses.

MAI Image 2.7 Flash Targets Faster and Cheaper Image Generation

Microsoft has also launched MAI Image 2.7 Flash, a smaller and faster variant designed for workloads where speed, cost and throughput are important.

The Flash model is aimed at businesses and developers that need to generate or edit large volumes of images, including creative testing, product visualisation, marketing assets, content production and automated image workflows.

Microsoft estimates the flagship MAI Image 2.7 costs approximately $35 per 1,000 images, while MAI Image 2.7 Flash costs around $10 per 1,000 images.

ModelApproximate cost per 1,000 imagesPrimary focus
MAI Image 2.7$35High-quality generation and precise editing
MAI Image 2.7 Flash$10Speed, efficiency and large-scale production

The pricing figures are Microsoft’s approximate estimates and can vary depending on usage and implementation.

The lower-cost Flash model could be particularly useful for companies producing thousands of product images, advertising variations, social media creatives, illustrations and visual assets.

Transparent Backgrounds Expand Commercial Image Use

Both MAI Image 2.7 models support transparent backgrounds, enabling users to generate isolated objects and visuals without requiring additional background-removal tools.

This feature is useful for creating product cutouts, logos, icons, illustrations, design elements and marketing materials that can be placed directly into other creative projects.

Transparent image generation can reduce the time required to prepare assets for websites, advertisements, e-commerce catalogues, presentations and digital campaigns.

For e-commerce businesses, it could help create product visuals that can be used across different backgrounds and promotional formats.

The capability also makes the models more useful for designers working with layered compositions, where individual objects need to be moved, resized or combined with other visual elements.

Microsoft Introduces Canvas for Interactive Image Editing

Microsoft is also introducing Canvas, an interactive image-creation and editing workspace powered by MAI Image 2.7 Flash.

Canvas allows users to start with a text prompt, sketch or existing image and refine the result through an interactive workflow.

Users can select specific areas of an image to replace, reposition or restyle rather than repeatedly generating an entirely new image.

The workspace is designed to support iterative creativity, allowing users to explore variations and refine their visual ideas in one environment.

The combination of image generation and interactive editing could make AI tools more useful for designers who need to maintain creative control rather than simply accept the first generated result.

MAI Image 2.7 Is Available Through Microsoft Foundry and Partners

Microsoft has made MAI Image 2.7 and MAI Image 2.7 Flash available in public preview through Microsoft Foundry.

The models are also being made available through partners including fal, Genspark, OpenRouter and Vercel.

Microsoft is expanding the distribution of its image models beyond its own products to reach developers, creative platforms, businesses and AI application builders.

MAI Image is also the model family powering Bing Image Creator, according to Microsoft, with the company reporting growing usage since its earlier rollout.

The broader partner strategy could help Microsoft compete in the image-generation market by making its models accessible across different platforms and creative workflows.

Microsoft Expands Its AI Model Portfolio Beyond Copilot

The MAI Image 2.7 launch is part of Microsoft’s broader strategy to develop specialised AI models that can serve different types of workloads.

The company is expanding its model portfolio across coding, cybersecurity, image generation, speech, transcription and other capabilities.

Developing in-house models could give Microsoft more control over model performance, pricing, infrastructure efficiency and integration with its cloud and software products.

It could also reduce Microsoft’s dependence on external AI model providers for specific workloads while offering customers a wider choice of AI systems.

For developers and enterprises, specialised models could provide a better balance between performance and cost depending on whether the workload involves image editing, coding, reasoning, audio processing or automation.

The Bigger Picture

Microsoft’s MAI Image 2.7 launch highlights how competition in generative AI is increasingly moving beyond simply creating impressive images.

Image editing precision, consistency across revisions, transparent backgrounds, production speed and cost per image are becoming important commercial differentiators.

The ability to change specific elements without disrupting an entire composition could make AI image generation more practical for businesses that need repeatable visual workflows.

Microsoft is also using its broader ecosystem, including Bing Image Creator, Microsoft Foundry and external partners, to increase distribution and adoption of its models.

As generative AI becomes more widely used by businesses, the companies that can combine high-quality results with predictable costs and useful editing controls may have an advantage in the commercial market.

Looking Ahead

MAI Image 2.7 and its Flash variant are expected to compete for adoption among designers, developers, marketers, e-commerce companies and businesses that generate images at scale. Microsoft will likely continue expanding model availability, improving editing capabilities and integrating its image-generation technology into additional creative and enterprise workflows. Adoption will depend on the models’ real-world performance, consistency, affordability and ability to deliver useful results with fewer editing iterations.

The launch also strengthens Microsoft’s ambition to build a broader portfolio of first-party AI models that can compete across creative and enterprise applications. As image generation becomes more integrated into marketing, product design, advertising and digital content production, the competitive focus will increasingly shift towards controllable editing, workflow integration and cost efficiency. MAI Image 2.7 represents Microsoft’s attempt to make advanced AI image creation more practical for everyday creative work and large-scale commercial production.

Get the day’s top stories in your inbox

One concise email. No spam, unsubscribe anytime.