ChatGPT Image Prompt Generator

This free image prompt generator writes prompts for ChatGPT image generation and OpenAI's GPT Image 2.5 Sunburst and Flare models. Describe the image you need, choose a size and style, and copy a prompt laid out the way OpenAI's image guide recommends.
Models last verified:

Generate your prompt

Access 40,000+ AI Prompts for ChatGPT, Claude & Codex — Free!

Subscribe to get instant access to our complete Notion Prompt Library — the largest curated collection of prompts for ChatGPT, Claude, OpenAI Codex, and other leading AI models. Optimized for real-world workflows across coding, research, content creation, and business.

Access Free Prompt Library

Which image model should I choose?

Compare GPT Image 2.5 Sunburst and GPT Image 2.5 Flare by best use, speed and default selection; both cost the same per token.

ModelBest useSpeed (OpenAI rating)Price per 1M tokens
GPT Image 2.5 SunburstMost capable model for image generation and editing; use where editing precision matters most.Not stated$5.00 text in / $30.00 image out
GPT Image 2.5 FlareFast, high-quality everyday image generation.Fastest$5.00 text in / $30.00 image out
From OpenAI’s model pages, checked 17 October 2026.

How do I use my prompt?

  1. Choose Create, Edit or Combine mode, then select Sunburst or Flare.
  2. Describe the subject, intended use, scene and style. Add exact text, editing instructions or labelled reference images if needed.
  3. Choose the image size, quality and format.
  4. Generate and copy the prompt or matching API request.
  5. Paste the prompt into ChatGPT, or use it in an OpenAI Images API or Responses request.

Prompting tips for GPT Image 2.5

  • Use quality: low for quick drafts, then compare higher quality settings for final assets to balance detail, latency, and cost. Image generation guide
  • Use the three recommended sizes when they fit; custom dimensions must obey the multiple-of-16, aspect-ratio, edge, and pixel-count limits. Image generation guide
  • For transparent output, request a transparent background and choose PNG or WebP rather than JPEG. Image generation guide
  • When editing with a mask, treat it as guidance rather than an exact boundary, and describe the desired change clearly. Image generation guide
  • When multiple images are supplied for a masked edit, make the intended edit image the first input because the mask applies there. Image generation guide
  • In Responses image generation, inspect revised_prompt so you can understand the automatic prompt revision applied by the mainline model. Image generation guide

Worked examples

Blog header

Before: Make a blog header about home coffee brewing.

After (made with this generator):

Use: blog header image for an article about home coffee brewing.
Scene: a sunlit kitchen counter in the morning, soft window light from the left.
Subject: a matte black coffee grinder beside a glass pour-over jug.
Details: steam rising from the jug, a few coffee beans on the counter.
Style: photorealistic, 35mm lens, shallow depth of field, natural colours.
Composition: subject on the right third; leave the left third clear for a headline.
Text: none.
Avoid: people, brand logos, clutter.

GPT Image 2.5 Flare, 1536x864 resized to 1200x675, quality high, JPEG.

What the generator added: The prompt adds a sunlit kitchen scene, specific coffee equipment, steam and scattered beans, plus photorealistic camera and lens details. It places the subject on the right, reserves headline space on the left and excludes text, people, logos and clutter.

Product background swap

Before: Replace the background in my white mug photo with a light-grey studio backdrop.

After (made with this generator):

Image 1: product photo of a white ceramic mug on a wooden table.
Change only the background: replace the wooden table with a plain light-grey studio backdrop and a soft shadow under the mug.
Keep everything else the same: the mug's shape, colour, logo, angle and lighting.

GPT Image 2.5 Sunburst, edit, same size as the original, PNG.

What the generator added: The prompt labels the source image and specifies a soft shadow under the mug. It tells the model to change only the background and keep the mug's shape, colour, logo, angle and lighting.

Sale poster with text

Before: Make a spring sale poster for a clothing shop with the words SPRING SALE.

After (made with this generator):

Use: poster for a spring sale in a clothing shop.
Subject: a pastel spring flower arrangement.
Style: flat vector illustration, clean shapes, limited colour palette.
Composition: flowers framing the edges, open space in the centre.
Text: "SPRING SALE" in bold sans serif, large, white, top centre.

GPT Image 2.5 Sunburst, 1024x1536, quality high, PNG.

What the generator added: The prompt specifies pastel flowers, flat vector shapes and a limited colour palette, with flowers framing an open centre. It quotes the exact headline and sets its font, size, colour and placement.

Frequently asked questions

Does this tool make images?

No. It writes the prompt; you paste it into ChatGPT or send it to OpenAI’s Images API.

Sunburst or Flare?

Sunburst for the best quality, edits and text in images; Flare for fast everyday images. Both cost the same.

How do I get readable text in an image?

Put the exact words in quotes, give the font, size, colour and position, and use medium quality or higher.

Can I edit a photo?

Yes. Use edit mode, say exactly what to change, and keep everything else the same.

Which sizes work?

The presets, or a custom size that follows OpenAI’s rules, which the form checks for you.