Edit photos with GPT Image 2 by just describing the change, free to tryTry It Free
ChatGPTImageEditor.com logoGPT Image Editor

Reference images (optional)

For the best result
  • Optional: upload an image to edit, plus up to 3 more as references. With no image, your prompt alone makes the picture
  • PNG, JPG or WEBP, up to 10 MB each
  • With a photo, say what to change and what to keep
0/5000
Aspect ratio

GPT Image 1.5

GPT Image 1.5 online: describe an image, or upload a photo and say what to change and what to keep. Your first image is free, no sign-up.

Original photo: Man in his early 40s: dressed in a navy suit in a bright officeBefore
GPT Image 1.5 result: Man in his early 40s: dressed in a navy suit in a bright officeAfter

Example result. Describe your image to make yours.

Examples

Original photo: Man in his early 40s: dressed in a navy suit in a bright officeBefore
GPT Image 1.5 result: Man in his early 40s: dressed in a navy suit in a bright officeAfter
GPT Image 1.5 result

What is GPT Image 1.5?

GPT Image 1.5 is the OpenAI image model announced in December 2025. OpenAI described it as making more precise edits that keep faces and logos, following instructions better and rendering text better than GPT Image 1.

This page runs GPT Image 1.5 as a photo editor and image creator. Upload the picture you want to change, add up to three reference images, and write what to change and what to keep, or skip the upload and describe a new image. You get one image in 1:1, 3:2 or 2:3 to download.

This is an independent site, not OpenAI and not ChatGPT. There is no chat: every request is one prompt and one result.

Made for

  • Photo edits
  • Product shots
  • Reference-based edits
  • Square and tall posts
  • Trying the older model

What to make with GPT Image 1.5

Six edits that fit a model built for precise changes.

Recolor a product

Upload a product photo and ask for another color or a cleaner background. Say what stays, such as the shape and the label, then compare the result with your original.

Edit with a reference

Add a second image, such as a jacket, a logo or a look to copy, and tell the model how to use it. You can add up to three references next to the photo you edit.

Put text in the picture

Ask for a sign, a label or a poster headline. OpenAI says GPT Image 1.5 renders text better than GPT Image 1. Check the spelling in the result before you use it.

Change the style

Turn a photo into a watercolor, a vintage print or a flat illustration. Name what the picture must keep, such as the pose or the composition.

Square, wide or tall

Pick 1:1 for a feed post or a profile picture, 3:2 for a wide shot or 2:3 for a tall one.

Small touch-ups

Ask to remove a stray object or change one detail, like a shirt color. A prompt is an instruction, not a lock, so look over the whole image.

GPT Image 1.5 vs GPT Image 2 and 2.5

How the model pages on this site differ. Not a quality ranking.

GPT Image 1.5GPT Image 2 and 2.5
Images per runUp to 4: the photo plus 3 referencesUp to 16: the photo plus 15 references
Result shape1:1, 3:2 or 2:3Auto or 7 ratios, at 1K
From OpenAIDecember 2025: more precise edits, better text than GPT Image 1GPT Image 2 added reasoning to image generation; 2.5 followed in September 2026
Free hereOne image for the whole site, no sign-upThe same shared allowance
Pick it whenA few references and a square, wide or tall result are enoughYou need more references or more ratios

Frequently asked questions

It is an OpenAI image model announced in December 2025. OpenAI said it makes more precise edits, follows instructions better and renders text better than GPT Image 1. This page runs it so you can try it on your own photo.

No. This is an independent site that runs the model, not OpenAI and not ChatGPT. You need no ChatGPT account, there is no chat, and there is no API for visitors. One prompt gives one image.

Yes, one image is free, with no sign-up and no credit card. The allowance is per visitor for the whole site, not per page. A run that fails is not counted. After that the tool says your free tries are used up, and there is no paid plan.

This page offers 1:1 square, 3:2 wide and 2:3 tall, with no resolution choice. For more ratios, use a GPT Image 2 page.

Up to 4 per run, or none: the image to edit plus up to 3 references. JPG, PNG, WEBP and iPhone HEIC photos work, up to 10 MB each. Large photos are shrunk in your browser before they upload.

Yes. Leave the upload empty and describe the image: GPT Image 1.5 creates it from your prompt. With a photo uploaded, it edits that photo instead.

GPT Image 2 is the newer model, and OpenAI added reasoning to image generation in it. Use this page for a few references and a 1:1, 3:2 or 2:3 result. Use the GPT Image 2 page for up to 16 images and more ratios. Your free image covers one run on the whole site, so pick before you start.

Say what to change and what to keep: this page sends your prompt to the model as written. It is AI output, so compare the result with your original, especially faces, logos and any text.

There is no chat, so the model does not remember the last result. Download your GPT Image 1.5 result to keep it. The single free image does not include another editing run.

Your photo is only used to create your images and is never used to train AI models. Uploads are stored on Cloudflare, and your prompt and images go to the model provider, which keeps the result for a limited time so you can download it. You must be 18 or older.

More tools