fal
@fal
MAI-Image-2.5 Image Editing is now available on fal! https://t.co/202nHAxmxp
8retweets39likes
View on Xfal adds image editing for the #2 ranked MAI-Image-2.5 model
fal has integrated image editing support for MAI-Image-2.5, a photorealistic image generation and editing model developed by Microsoft. This update follows the platform's recent MAI-Image-2.5 integration, now adding a dedicated endpoint for image-to-image transformations. The system uses inference (running a trained model to generate outputs) to modify uploaded images based on natural language prompts.
- Model
- MAI-Image-2.5
- Output Price
- $0.05 per image
- Input Text Price
- $5.00 per 1M tokens
- Input Image Price
- $8.00 per 1M tokens
The addition is notable as MAI-Image-2.5 currently holds the second position on the Arena.ai Image Edit leaderboard. Built for commercial use, the model turns text prompts or uploaded images into high-quality, design-ready visuals with fine-grained, pixel-level control. It offers a high-fidelity option for professional workflows that need precise edits to existing assets.
Teams can access the editing endpoint through the fal API or its web playground. Generation costs $0.05 per image, with text input billed at $5.00 per million tokens and image input at $8.00 per million tokens. The endpoint applies localized, prompt-driven edits with the fine-grained, pixel-level control fal cites for design-ready output.
Still wondering? A few quick answers below.
MAI-Image-2.5 is Microsoft's photorealistic image generation and editing model. It turns text prompts or uploaded images into high-quality, design-ready visuals with fine-grained, pixel-level control, and is offered for commercial use on fal.
Image editing using this model on fal costs approximately $0.05 per generated image. In addition to the per-image fee, users are billed for input tokens at a rate of $5.00 per 1 million text tokens and $8.00 per 1 million image input tokens. Pricing is subject to change.
This MAI-Image-2.5 editing endpoint is built for commercial use and offers fine-grained, pixel-level control over existing images. The model currently ranks second on the Arena.ai Image Edit leaderboard, and it applies localized, prompt-driven modifications to source images for design-ready results.
Every HeadsUpAI update is written based on its original source and reviewed before it's published. Read our editorial standards →
