CheaperGen
OpenAITutorial

How to Use GPT Image 2.5

Make an GPT Image 2.5 image in the CheaperGen Studio — no code, no install, no OpenAI account. Pick a mode, write one clear prompt, set the size and press Generate. A test costs $0.007; developers can follow the API tutorial instead.

Modes
2
Text → Image · Edit
Test image
$0.007
1K
Sizes
up to 4K
3 options
Failed runs
$0
refunded automatically
Create with GPT Image 2.5Get API keyAPI docs

Prompt builder

Build a GPT Image 2.5 prompt

Fill in the parts of the shot; the prompt assembles itself. Send it to the playground below and run it — the steps after it explain each part.

Your GPT Image 2.5 prompt

Fill in a few boxes — or start from the example — and the prompt appears here.

Try it

Try GPT Image 2.5 here

Runs on your balance at the price on the button; failed runs are refunded. Results are saved to your account.

Open the full-screen Studio

Before you start

Which GPT Image 2.5 mode to use

GPT Image 2.5 has 2 modes. Pick by what you already have.

GPT Image 2.5 modes
ModeYou give itBest forFrom
Text → ImageA promptImages from a description: concepts, illustrations, ad visuals.$0.0061/image
EditUp to 16 images and an instructionChanging an image you have: backgrounds, colours, outfits, or combining several images.$0.0061/image

Step by step

How to use GPT Image 2.5 in 6 steps

  1. 1

    Open GPT Image 2.5 in the Studio

    Open the Studio (it opens with GPT Image 2.5 selected), or choose the Image tab and pick GPT Image 2.5 in the model menu. Sign in with Google or your email; you pay per generation from prepaid credits.

  2. 2

    Choose what to make

    • Text → Image — a prompt.
    • Edit — up to 16 images and an instruction.
  3. 3

    Write the prompt

    One shot per prompt works best. Cover these, in plain sentences:

    • Subject. The main subject and its key details: material, colour, expression, product features.
    • Composition. Framing and angle: close-up, flat lay, three-quarter view, rule of thirds, negative space.
    • Setting and light. Background, environment and lighting: studio softbox, golden hour, neon.
    • Style. Photo, illustration, 3D render or painting — plus lens or medium details.
    • Edit instruction. For edits, say what to change and what must stay the same: “replace the background, keep the product and its shadow”.

    More structure and examples: the GPT Image 2.5 prompt guide.

  4. 4

    Set the options

    The settings sit under the prompt; rarely needed ones are behind ⋯. What each does — and to the price — is in the table below.

  5. 5

    Generate

    The exact price is shown before you run: press Generate (or ⌘/Ctrl + Enter). The amount is held from your balance and refunded in full if the generation fails.

  6. 6

    Download and iterate

    Results appear above the prompt; download them from there. Change one thing per try — the prompt, the composition or a setting — so you know what helped. Compare runs the same prompt on several models side by side, and Code shows the same request for the API.

Settings

GPT Image 2.5 settings explained

For Text → Image; the other modes add their inputs (frames, references or images to edit). Prices are per image.

GPT Image 2.5 settings
SettingChoicesDefaultWhat it doesPrice
Aspect ratio1:1, 3:2, 2:3, 4:3, 3:4, 5:4, 4:5, 16:9, 9:16, 21:9, 9:211:1Output image shape.Changes the price
Resolution1K, 2K, 4K1KOutput size tier: 1K (≈1 MP), 2K (≈4 MP) or 4K (≈8 MP, experimental).$0.0061–$0.0114/image
QualityLow, Medium, High, X-High, MaxMediumRendering effort. Higher quality uses more output tokens and costs more; xhigh and max are for final assets.$0.0061–$0.0492/image

Every tier and a calculator: GPT Image 2.5 API Pricing.

Worked example

A first prompt to try

Text → Image with the default settings. Paste it into the Studio and change one thing at a time.

“Packaging design for an oat-milk carton called “OATLY GOOD”, pastel illustration, nutrition panel with legible text”

Aspect ratio: 1:1 · Resolution: 1K · Quality: Medium

$0.013$0.014 for this image

Try it in the Studio

Avoid these

Common GPT Image 2.5 mistakes

Drafting at 2K · high

Every try at 2K · high costs 16.5× one at 1K · low. Settle the composition and motion at 1K · low, then render the keeper at the size you deliver.

Vague edit instructions

Say what to change and what must stay: "replace the background with a beach at sunset, keep the product and its shadow".

Describing what you don't want

Prompts work better when they describe what should be on screen. "No blur" or "no extra fingers" can pull the result towards the very thing you name.

FAQ

How to use GPT Image 2.5: questions

Yes. Open it in the CheaperGen Studio, pick a mode, write a prompt and press Generate. No code, no install and no OpenAI account.

There is no free tier, and no subscription is needed either: you pay per image, from $0.007 for a 1K · low image — up to 8% below the official price. Credits are prepaid and failed generations are refunded. Every price is on the GPT Image 2.5 pricing page.

Yes: as images to edit. Upload them in the Studio, or pass their URLs in the API.

Write one clear shot per prompt — subject, action, setting, camera and light — and change one thing per try. The GPT Image 2.5 prompt guide has a structure and examples that work.

Yes. The same model and price are on the API — see how to use the GPT Image 2.5 API — and on our MCP server, so agents like Claude can generate for you (MCP setup).

Nothing is charged. The price is held when you press Generate and refunded in full if the generation fails or you cancel it while it is queued.

It depends on quality, speed and price; the GPT Image API page compares all 2 side by side.

GPT Image 2.5 is developed by OpenAI. CheaperGen provides API access and does not own the model. Trademarks belong to their respective owners.