When to use Image to Image
Use Image to Image when you already have one or more photos or renders and want to change style, swap backgrounds, combine elements, or refine details.
Tagline: “Add reference photos, describe what you want changed or combined, then generate a new image from your prompt.”
Reference image rules
You must upload at least one reference image before generating. You can add up to eight images.
Default maximum file size is 12 MB per image. GPT Image 2 references must be JPEG, PNG, or WebP and are limited to 10 MB each.
- Use sharp, well-lit sources when possible
- Tell the model what to preserve (“keep product label readable”)
- Tell the model what to change (“replace background with sunset beach”)
- GPT Image 2: only JPEG, PNG, or WebP uploads are accepted
Workflow
Open Dashboard → Image to Image, then:
- Click Add images and upload references
- Write a change-focused prompt (up to 4,000 characters; counter soft-warns above 1,000)
- Choose aspect ratio and quality tier
- Pick model and optional style in Advanced
- Generate and compare against the reference in Creations
Get better edits with fewer retries
Smaller, specific edits succeed more often than “make it completely different.”
When combining multiple references, name the role of each image in the prompt (e.g., “use figure from image 1, palette from image 2”).
Category: Image generation