How much does one image cost?

It depends on the settings. The exact price is on the button before you start, and it updates with every change.

What if it doesn't work out?

If the run fails at the provider, no credits are charged. If the result is ready but you don't like it, press “Edit and retry”: the prompt and settings stay.

Which frame shapes are there?

1:1, 2:3, 3:2, 4:3, 3:4, 4:5, 5:4, 16:9, 9:16, 21:9. 9:16 suits Reels, Shorts and stories; 16:9 suits YouTube and websites.

Can I use my own picture?

Yes: add up to 16 — from your computer, from the Library or as an element. The model takes them as a base or a reference.

How is GPT Image 2 different from GPT Image 2.5 Sunburst?

Both are in the OpenAI line-up. Switch between them in the line-up at the top of the page — the price of each is on its card.

Where are the results kept?

Every finished result goes to your Library. From there you can download it, publish it in the Community or use it as the base for the next one.

GPT Image 2Top

OpenAI · Image · 1K–4K

API

Saved

Example

Your picture will appear here

Fill in the settings and press Run. The result is also saved to your library.

Price

Reference price

No published price yet - this model is not in the public price list.

What the price depends on

This model is billed per generation, so the price does not depend on how long you wait.
Resolution and quality settings change the price. Change one and the number above follows.
A generation the provider fails is not charged. A chat answer is charged for what the model wrote and thought, even if you stop it.

Prices for every model Top up

Prices are in credits. What a credit costs is on the pricing page.

Which one to pick

Model	Key specs	Price from
GPT Image 2.5 SunburstNew	1K–4K	—
GPT Image 2Top· you are here	1K–4K	—
GPT Image 1.5	1024x1024, 1024x1536, 1536x1024	—
GPT-4o Image	1:1, 2:3, 3:2	—

About the model

OpenAI GPT Image 2: three quality levels, resolution up to 4K and up to 16 reference images.

Produces: Image
Takes: text, pictures
Prompt: 20,000 characters
Frame shape: 1:1, 2:3, 3:2, 4:3, 3:4, 4:5, 5:4, 16:9, 9:16, 21:9
Resolution: 1K, 2K, 4K

What you can set

Every setting of this model and the choices it allows — the same as in the form.

Input images
up to 16 files
Quality
DraftStandardHigh
Resolution
1K2K4K
Aspect ratio
Auto1:12:33:24:33:44:55:416:99:1621:9

For developers

Run this from your own code

The same request the Generate tab sends, with the settings you have chosen.

Model id:gpt-image-2

curl -X POST https://seedara.ai/api/v1/generate \
  -H "Authorization: Bearer $SEEDARA_API_KEY" \
  -H "Content-Type: application/json" \
  --data '{
  "model": "gpt-image-2",
  "prompt": "Describe the image: subject, style, framing, lighting...",
  "params": {
    "quality": "low",
    "resolution": "1K",
    "size": "auto"
  }
}'

This snippet updates as you change the form. Copy it once the settings are the ones you want.

Request parameters

What goes in `params`, by the names the API uses.

Optional

Left out unless you set them - the provider then applies its own default.

Parameter	Type	Accepted values	Default
`image_urls` Input images	images	up to 16 files	—
`quality` Quality	one of a list	`lowmediumhigh`	low
`resolution` Resolution	one of a list	`1K2K4K`	1K
`size` Aspect ratio	one of a list	`auto1:12:33:24:33:44:55:416:99:1621:9`	auto

How a generation runs

POST /v1/generate answers immediately with a job_id - it does not wait for the result.
Poll GET /v1/jobs/{id} until status is completed or failed.
A completed job carries output_url - the finished file.

curl https://seedara.ai/api/v1/jobs/JOB_ID \
  -H "Authorization: Bearer $SEEDARA_API_KEY"

Questions

Frequently asked about GPT Image 2

New

Start now

Your first image —
in a minute

Describe what you want to see — the model does the rest.

Takes you to the settings — prompt, format and price are already there

No auto-charges
Pay only for a finished result
Works are kept in your Library

100credits as a gift

1:1 · 2:3 · 3:2Frame shape