OPENAI · AI IMAGE GENERATOR
Describe the image, or add up to 15 reference images to guide it. GPT Image 2 follows layout instructions closely — titles, numbered steps, labels and interface elements — with 1K, 2K and 4K output.
Review results, manage assets and access generation history in AI Studio.
Every example below was actually generated with GPT Image 2 on ImgAI, with the original prompt and settings. Click "Use this example" to load it into the creator.
GPT Image 2 · 1K · 2:3
Title, subtitle and footer all landed where the prompt placed them, with clean margins.
A minimalist travel poster for Kyoto in autumn: a single red maple branch over a quiet temple roof, warm cream background. Large title "KYOTO", subtitle "Autumn Leaves Festival", small footer "November 2027". Clean modern typography with generous margins.
GPT Image 2 · 1K · 1:1
Soft side light and shallow depth of field. The model added the slogan on the mug by itself — the prompt did not ask for any text.
Studio product photo of a matte black ceramic coffee mug on a light oak table, soft window light from the left, a few roasted coffee beans beside it, shallow depth of field, clean e-commerce look.
GPT Image 2 · 1K · 3:4
Four numbered steps, each with an icon and a one-line caption, in a warm brown palette.
A clean flat infographic titled "How Coffee Is Made" with four numbered steps: 1 Harvest, 2 Dry, 3 Roast, 4 Brew. Each step has a simple icon and a one-line caption. Warm brown and cream palette, white background.
GPT Image 2 · 1K · 2:3
A flat UI mockup with the logo, fields, button and link requested in the prompt.
A mobile app sign-in screen for a fictional meditation app called "Calma": soft lavender gradient, a small lotus logo, an email field, a password field, a "Sign in" button and a "Create account" link. Flat UI design, no device frame.
GPT Image 2 · 1K · 1:1
A gouache-textured children's book illustration in soft greens and oranges.
Children's storybook illustration of a small fox reading a book under a mushroom during light rain, gouache texture, soft greens and oranges, cozy and gentle.
GPT Image 2 is an image generation and editing model from OpenAI. It creates images from a text prompt and can also use reference images as a starting point. Its strength is putting what you explicitly ask for onto the canvas: a title in a given position, four numbered steps with captions, a sign-in screen with the fields you listed.
On ImgAI you can use it with up to 15 reference images, choose 1K, 2K or 4K output and 15 aspect ratios from 3:1 to 1:3, and write prompts of up to 10,000 characters. One image costs 15–35 credits depending on resolution. In our own tests it sometimes added short text we did not ask for (for example a slogan printed on a mug), so check the text in the result before you publish it.
Based on the model's published capabilities and what we actually generated, to help you judge whether it fits your task.
Titles, subtitles and dates set where you specify them. Good for travel posters, event announcements and social covers.
Numbered steps, icons and one-line captions in a clean layout — the kind of image that falls apart when text is garbled.
Studio-style shots with controlled light and props. Add your real product as a reference image to keep its shape and label.
Flat app screens with the fields and buttons you name, useful for pitches and early design discussions.
Specs come from the configuration currently available on this site; the creator always shows the live price. Failed generations are refunded automatically.
| Input | Text + up to 15 reference images |
| Resolution | 1K · 2K · 4K |
| Prompt length limit | 10,000 characters |
Supported aspect ratios
| Tier | 1K | 2K | 4K |
|---|---|---|---|
| GPT Image 2 | 15 | 25 | 35 |
These come from real generations — read them alongside the original prompts in the examples.
Write the exact text, e.g. title "KYOTO", subtitle "Autumn Leaves Festival". Unquoted descriptions leave the wording up to the model.
Say what is large, what is small and where it sits: "large title at the top, small footer at the bottom, generous margins".
If the product should stay blank, write "no text or logo on the mug". The model otherwise may add a slogan of its own.
1K is the cheapest tier. Iterate there until the layout is right, then rerun the same prompt at 2K or 4K for the final file.
GPT Image 2 is OpenAI's image generation and editing model. On ImgAI it takes a text prompt plus up to 15 reference images and outputs 1K, 2K or 4K images in 15 aspect ratios.
Each image costs 15–35 credits depending on the resolution you choose (1K, 2K or 4K). The creator shows the exact price before you generate, and failed generations are refunded automatically.
GPT Image 2.5 is the newer model and has its own page on ImgAI, with quality tiers and up to 16 reference images. GPT Image 2 costs fewer credits per image and remains a solid choice for posters, infographics and product shots.
Yes. Upload the image as a reference and describe the change — for example "replace the background with a white studio wall". You can add up to 15 reference images in one request.
Text accuracy drops with long paragraphs and very small type. Keep each text element short, put it in quotes, and check spelling in the result. Raising the resolution helps small text stay legible.
Yes. GPT Image 2 accepts prompts in other languages, and text you put in quotes is rendered in that language. For non-Latin scripts, keep the quoted text short and check it in the result.
Sign up and generate with credits. Results are saved to your library, ready to download or keep editing.