GPT Image is OpenAI's family of image generation models. On this page you can use GPT Image 1.5, GPT Image 2 and the GPT Image 2.5 versions to create images from text and to edit or combine your own images.

Which GPT Image version should I choose?

For most work, start with GPT Image 2.5 Flare: according to OpenAI it produces higher-quality images than GPT Image 2 at lower latency. Choose 2.5 Sunburst for demanding edits that need tighter control, GPT Image 2 as a proven all-rounder, or 1.5 for cheaper drafts. Some versions require a paid plan; credit costs are shown in the model selector.

Can I edit my own photos with GPT Image?

Yes. Switch to image editing, upload one or more images and describe the change you want. GPT Image will keep the parts you don't mention as close to the original as possible.

How many credits does GPT Image cost?

Each generation uses credits based on the version you pick. Current prices are listed next to every model on this page and on the pricing page. New users receive free credits to try it.

Can I use GPT Image results commercially?

You can use the images you generate in personal and commercial projects, subject to our Terms of Service and OpenAI's usage policies.

Create with GPT Image online

OpenAI's GPT Image models follow long, detailed prompts closely and render legible text inside images — posters, infographics, UI screens and photoreal scenes. Generate from text, edit a photo, or blend several references, no setup required.

Start Creating Pricing

Free credits on sign-up · No credit card required

Model

GPT Image 22 credits
GPT Image 2.5 Flare2 credits
GPT Image 2.5 Sunburst3 credits
GPT Image 1.51 credits

Model Selection

OpenAI GPT Image 2 model with enhanced image generation and editing capabilities

Prompt

⌘Enter0/1000

New accounts get 2 free credits · 2 credit per generation

Template Management

Generation Results

No generation results yet

Select a mode and enter a prompt to start creating

Made with GPT Image

Every image below was generated with GPT Image on this site. Open one to see the exact prompt and try it yourself.

New tea launch poster

GPT Image 2

Design a 3:4 vertical poster for a new Chinese trendy tea launch. Use a New Chinese visual style that feels light-luxury and restrained. The palette should be dark green, off-white, and gold, with rice-paper texture, elegant negative space, landscape accents, and modern layout design. Main subject: a visually appealing cold-brew tea with tea leaves, citrus, ice cubes, and touches of gold foil. The poster must accurately display the following exact Chinese copy: "山川茶事" "山柚观音" "冷泡系列" "新品上市" "一口清醒，半城入夏" "限定尝鲜价" "中杯 16 元" "大杯 19 元" "第二杯半价" "推荐风味" "观音茶底 / 西柚果香 / 轻乳云顶 / 冰感回甘" "扫码点单" "SHANCHUAN TEA" Maintain a clear promotional hierarchy while keeping the overall feeling sophisticated rather than cheap. Pay special attention to small text, numbers, prices and Chinese typography aesthetics.

Use this prompt

1/6

GPT Image: when the image has to be right

GPT Image is built for images that need to be accurate and usable — posters with real copy, infographics, UI mockups and product visuals. Describe it in detail and it follows the brief.

Readable text: Headlines, prices and labels render cleanly, in English and many other languages.
Structured layouts: Infographics, menus, slides and app screens keep their hierarchy and alignment.
Photoreal scenes: Natural light, reflections and everyday details make photos believable.

GPT Image: when the image has to be right

What GPT Image does well

The GPT Image family is built for prompts that describe exactly what you want

Follows complex prompts

Describe layout, objects, lighting and mood in plain language. GPT Image keeps track of many details in a single prompt.

Readable text in images

Posters, menus, labels and UI mockups with text that actually reads correctly — one of GPT Image's best-known strengths.

Edit with words

Upload a photo and say what to change: swap the background, restyle the scene or add and remove objects.

Multiple references

Combine up to 5 reference images (depending on the version) to keep products, characters or styles consistent.

Sizes and aspect ratios

Pick the aspect ratio and output size that fits your use, from square social posts to wide banners.

Four versions to choose from

GPT Image 2.5 Flare is the fast everyday choice with higher quality than GPT Image 2; 2.5 Sunburst is tuned for demanding, tightly controlled edits. GPT Image 2 and 1.5 are also available.

What people make with GPT Image

From marketing assets to everyday visuals — wherever text and layout matter

Posters & ads
Launch posters, promo banners and social ads with accurate copy and prices.
Infographics
Explainers, process diagrams and knowledge cards that are easy to read.
UI mockups
App screens and web pages for pitches, prototypes and presentations.
Product images
Clean product shots and detail-page visuals for online stores.
Menus & signage
Café menus, price boards and in-store signs with exact wording.
Realistic photos
Street, lifestyle and documentary scenes that look captured, not generated.
Teaching materials
Diagrams, flashcards and slides for classes and training.
Social content
Covers, carousels and thumbnails ready to post.

Edit and combine with plain language

Upload one or more images and describe the change. GPT Image keeps what you don't mention and changes what you do — from the season in a street photo to the background of a product shot.

Targeted edits: Swap backgrounds, change the weather or restyle a scene while the composition stays put.
Multiple references: Combine up to 5 images (depending on the version) to keep products and characters consistent.
Pick the right version: Use 2.5 Flare for speed, 2.5 Sunburst for demanding edits, or GPT Image 2 and 1.5.