On the Firefly homepage, select Generate in the left panel and then Generate image.
Last updated on
9 June 2026
Learn how to use GPT Image models to generate image variations using simple text descriptions.
Use GPT Image models such as GPT Image and GPT Image 1.5 with the Generate image feature to create images from text prompts in Firefly. These models can help generate detailed visuals across a range of creative styles and concepts. Adjust your prompts and Firefly settings to refine generated outputs for your projects.
When you select a partner model to generate images, you'll be prompted to confirm that you're using a model that isn't developed by Adobe, along with other relevant information. Please review the details and select OK to continue.
On the Generate image page, enter a text prompt in the Prompt field.
In the General settings section, select the Model dropdown menu.
Under the Partner models section, select GPT Image or GPT Image 1.5.
Use the Aspect ratio dropdown menu to select from the following options:
- Wide (4:3)
- Square (1:1)
- Tall (2:3)
When using GPT Image as the model, use the Subject option to upload an image to edit, such as adding an element to it.
The example below shows how the uploaded image, combined with a text prompt, generates an image with a scarf added.
Uploaded image
Generated image
When using GPT Image 1.5 as the model, use the Reference images option to upload up to six images to share as a reference or edit them, such as adding an element to them.
Select Generate .
Turn text into art with Adobe Firefly
Create beautiful images and video clips from text prompts using generative AI.
