HeadsUpAI

fal adds Microsoft MAI-Image-2.5 for photorealistic commercial design and text

fal has added support for MAI-Image-2.5, a photorealistic image generation and editing model from Microsoft. It produces high-quality visuals from text prompts or uploaded images with pixel-level control. The model specifically targets improvements in natural lighting, accurate skin tones, and legible text rendering for commercial applications like branding and packaging.
Model
MAI-Image-2.5
Pricing (Output)
$0.05 per image
Pricing (Input)
$5.00 per 1M tokens
Key Capabilities
Text rendering and photorealism
Use Cases
Branding and packaging

The integration makes the model available for serverless inference (running a trained model to generate outputs) on the fal infrastructure. This release follows the model's recent No. 2 debut for image editing on the Arena leaderboard. The model is also available through other providers, including OpenRouter.

Access is available via the fal API or playground for text-to-image and image-editing tasks. Generation costs approximately $0.05 per image, with input text tokens billed at $5.00 per million—billed per image and rounded up to the nearest $0.0001. This makes high-fidelity, design-ready visuals accessible directly within production inference workflows.

View the full update on fal.ai
fal
fal
@fal
X

🚨 MAI-Image-2.5 is now live on fal! 📸 Photorealistic images with natural lighting and accurate skin tones Refined text rendering for branding, packaging, and commercial design Text-to-image and image editing with precise, design-ready control https://t.co/0HjrD4lTZY

10retweets70likes
View on X

Still wondering? A few quick answers below.

MAI-Image-2.5 is a high-fidelity image generation and editing model developed by Microsoft. It is specifically optimized for photorealistic visuals, natural lighting, and accurate skin tones. The model is designed for professional use cases where precise control over branding, packaging, and commercial design elements is required.

Using the model on the fal platform costs approximately $0.05 per generated image for output. Additionally, input text tokens are billed at a rate of $5.00 per million tokens. Each requested image is billed separately, and the total cost is rounded up to the nearest $0.0001.

The model features advanced text rendering for commercial design, natural lighting, and accurate skin tones. It supports both text-to-image generation and image-to-image editing with pixel-level control. These capabilities allow for the creation of design-ready visuals from simple text prompts or uploaded images.

Every HeadsUpAI update is written based on its original source and reviewed before it's published. Read our editorial standards →

Share this update