Microsoft has introduced MAI-Image-2.5-Pro, a new artificial intelligence (AI) image generation model designed to deliver higher-quality visuals, more accurate text rendering, and advanced image editing capabilities. The launch is part of Microsoft’s broader strategy to expand its portfolio of in-house AI models and reduce its dependence on third-party providers for AI-powered features across products such as Microsoft 365, PowerPoint, and Bing Image Creator.
MAI-Image-2.5-Pro builds on the MAI-Image-2.5 model introduced earlier this year and focuses on enterprise-grade image generation, offering improvements in photorealism, prompt adherence, and controllable editing. Microsoft says the model is optimized for commercial content creation, product imagery, marketing assets, and presentation graphics while maintaining competitive pricing for developers through Azure AI Foundry.
Microsoft Unveils MAI-Image-2.5-Pro
The new model supports both:
- Text-to-image generation.
- Image-to-image editing.
Microsoft says MAI-Image-2.5-Pro is intended for users who prioritize image quality over generation speed.
Model Overview
| Feature | Details |
|---|---|
| Model | MAI-Image-2.5-Pro |
| Developer | Microsoft |
| Primary Functions | Text-to-image generation and image editing |
| Platform | Azure AI Foundry |
| Target Users | Developers, enterprises, content creators |
Key Improvements
Microsoft highlights several upgrades over previous versions of its image model.
Better Image Quality
The company says MAI-Image-2.5-Pro produces:
- High-definition portraits.
- More realistic lighting.
- Improved material textures.
- Better reflections and shadows.
- Stronger scene composition from complex prompts.
Improved Text Rendering
One of the biggest improvements is text generation inside images.
The model is designed to produce:
- More accurate words on posters.
- Clear product labels.
- Better typography in marketing material.
- Improved packaging mockups.
Microsoft has positioned text rendering as one of the model’s strongest capabilities compared with earlier releases.
Advanced Image Editing
The model also supports precise editing of existing images, including:
- Object replacement.
- Background modification.
- Removal of unwanted elements.
- Targeted edits without changing unrelated portions of an image.
- Better preservation of facial identity across edits.
Technical Specifications
MAI-Image-2.5-Pro includes several enterprise-focused capabilities.
Technical Snapshot
| Specification | Details |
|---|---|
| Maximum Prompt Length | 32,000 tokens |
| Image Input Formats | JPEG, PNG |
| Maximum Output Resolution | 1,024 × 1,024 pixels (about 1 megapixel) |
| Prompt Language | Optimized for English |
| Text Rendering | English |
The model accepts both text prompts and existing images as inputs, enabling users to generate new visuals or edit existing ones.
Integration Across Microsoft’s Products
Microsoft has already begun deploying the MAI image family across its software ecosystem.
Examples include:
- Bing Image Creator, where MAI-Image-2.5 has become the default image generation model.
- PowerPoint, enabling users to generate and edit presentation graphics directly.
- OneDrive, where MAI-Image-2.5 powers AI-assisted image editing.
According to Microsoft, using its in-house model can reduce image-generation costs by up to 84% compared with relying on OpenAI’s GPT-Image-2 for certain workloads.
Pricing
Microsoft has published API pricing for developers using MAI-Image-2.5-Pro through Azure AI Foundry.
API Pricing
| Input/Output | Price (per 1 million tokens) |
|---|---|
| Text Input | $5 |
| Image Input | $8 |
| Image Output | $106 |
Developers can access the model through Azure AI Foundry’s Model Catalog for production applications.
Microsoft’s Expanding MAI Strategy
The release reflects Microsoft’s growing investment in proprietary AI models.
The MAI family now includes:
- MAI-Image-2.5-Pro.
- MAI-Image-2.5.
- MAI-Thinking reasoning models.
- MAI-Voice models.
- Coding-focused MAI models.
By developing its own multimodal AI stack, Microsoft aims to strengthen control over performance, pricing, and product integration while reducing reliance on external AI providers.
Looking Ahead
The launch of MAI-Image-2.5-Pro marks another step in Microsoft’s effort to build a comprehensive portfolio of in-house generative AI models spanning text, images, voice, and reasoning. With enhanced image quality, stronger text rendering, advanced editing capabilities, and integration across Microsoft 365, Bing, and Azure AI Foundry, the company is positioning the model as a premium solution for enterprise content creation and commercial design workflows.
Looking ahead, Microsoft’s continued investment in the MAI family signals intensifying competition in the AI image generation market, where providers are racing to improve realism, editing precision, and cost efficiency. As enterprises increasingly adopt generative AI for marketing, design, and productivity applications, proprietary models such as MAI-Image-2.5-Pro are expected to play a central role in Microsoft’s broader AI ecosystem and cloud strategy.
Get the day’s top stories in your inbox
One concise email. No spam, unsubscribe anytime.
