Change one detail. Keep the rest of the idea.
Create images with GPT Image 2.6 Flare or Sunburst. Write a prompt, upload references for an edit, and choose the quality, resolution and output format.
Select Model
0/5000 characters
A premium studio product photograph of a cobalt blue ceramic mug with a curved handle on a pale limestone table. One small bright orange rests beside the mug. Soft window light from the left, detailed ceramic glaze, quiet ivory background, no text, no logos. Square composition.
No Results Yet
Enter a prompt and click Generate to create your first GPT Image 2 image.
This workspace uses GPT Image 2.6 through NanoAPI. Flare is selected by default.
WHAT IS GPT IMAGE 2.6?
GPT Image 2.6 is available here through NanoAPI, with Flare and Sunburst variants. Generate from text or edit uploaded references, then review and download your result.
FROM BRIEF TO IMAGE
Use these examples to turn a GPT Image 2.6 idea into an inspectable brief. State what must change, what must stay and what you will check in the result.
01POSTERS & WORDING
For a city guide or campaign poster, supply exact words, their order and the space around them. Ask for a title that can be read before the smaller supporting text.
Compare the rendered spelling and punctuation with your original brief.
02PRODUCT DETAILS
For a watch or crafted object, specify the point of contact, the object being held and the direction of the light. Inspect fingers, reflections and tiny marks before using the image in a product story.
A convincing overall image still needs a close check of its contact points.
03LOCAL REFERENCE EDITS
A storefront color test should not become a different shop. Name the door or facade you want recolored, then list the windows, lettering, camera angle and materials that should stay consistent.
Upload a reference in edit mode; words alone do not attach an image.
MADE FOR YOUR NEXT TASK
Each task has a different review question: does the edit preserve the place, does the copy read, and does the product detail make sense?
SHOP OWNERS
Explore a door color against the existing facade, then discuss the result with the people who know the shop.
MAKERS & PRODUCT TEAMS
Plan close-up imagery around the hands, tools and material details that a customer needs to see.
CAMPAIGN DESIGNERS
Test the relationship between a short title, supporting copy and a strong image before setting the final layout.
THREE STEPS TO GET STARTED
Choose GPT Image 2.6 settings, describe your image and review the result.
Choose Flare or Sunburst. Use text-to-image for a new scene or image editing with your uploaded references.
Write the subject, composition, exact text and details to preserve. Add a reference for editing, then review the settings and credit estimate before generating.
Open the example at full size and compare its text, objects and layout with the brief. When your own image is complete, review it before downloading or continuing to edit.
Choose a plan or credit pack, with credits shown before checkout.
Perfect for individuals and light users
Billed annually at $90.00
Save 50%INCLUDES
For professional creators and teams
Billed annually at $234.00
Save 50%INCLUDES
Designed for large enterprises and professional studios
Billed annually at $960.00
Save 50%INCLUDES
To cancel your subscription, please contact support: support@gpt-image2.art
BEFORE YOU CREATE
Yes. This page connects to GPT Image 2.6 through NanoAPI. Choose Flare or Sunburst, write a prompt and generate. Upload references to edit an existing image.
The default Medium / 1K setting costs 8 credits per image. Other quality and resolution tiers have different prices. The total updates with your settings and image count before you submit.
Text-to-image starts from a written prompt. Reference editing needs an uploaded image and a description of the change. Separate the intended edit from the details you want to preserve.
The upload control accepts JPEG, PNG and WebP. The image count depends on the selected model and is shown beside the upload control. Use a clear reference that you have permission to upload.
Choose PNG, JPEG, WebP, GIF or SVG. GIF and SVG use 1K; SVG does not support reference images. GIF uses a custom size and animation settings. JPEG does not support a transparent background.
Open task history to inspect the status and error. Failed charged tasks follow the existing credit-return process; partial results are handled according to the task record. A refund is confirmed when its status is recorded. Contact support if the task and balance disagree.
Use text instructions and reference images to describe the area to change. This generator does not have a drawing canvas. You can upload an image that already contains your annotations.
START WITH ONE CLEAR IDEA
Choose a connected model, name one change and list the details that should stay. Review the image against that list before downloading it.