CreateForge
Image · Available

OpenAI image creation and editing

GPT Image 2 AI image generator

Create or edit images with GPT Image 2 using detailed prompts, private references, flexible framing, and resolution-aware protected quotes.

Provider
OpenAI
Starting quote
3 credits

GPT Image 2

0/20000
Aspect RatioAuto
Resolution
Output Image Count

Add a prompt to continue.

Current task & results

Completed results are saved to your Library

Start from an example

Use a prompt, then add your own references.

Choose a starting point

01 / 03

Use GPT Image 2 for creation or a controlled edit

Text-to-image is appropriate when composition and subject appearance can be invented from a brief. Reference editing is stronger when the source contains a property that must survive, such as product shape, face, camera, pose, or layout.

Reference files do not explain their own purpose. Name each role in the prompt, identify what remains fixed, and define the permitted change. A small coherent reference set usually communicates more clearly than many conflicting images.

Output planning

02 / 03

Choose frame and resolution for the delivery stage

Select an aspect ratio based on the final channel and write a composition designed for it. A vertical image needs a different subject position and depth structure from a wide campaign header; a later crop is not always an equivalent substitute.

Use an appropriate review resolution to prove the concept, then choose a higher delivery setting when fine detail is genuinely required. The protected quote in the generator is authoritative for the selected request.

Production review

03 / 03

Inspect the image before it becomes a public asset

Compare the saved result with the acceptance list: identity, object count, geometry, text, material, shadow, background continuity, and crop. Generated content can be non-unique or inaccurate even when it is visually polished.

Keep the final prompt and controls with the Library asset. If one property fails, revise that instruction while preserving the parts that already work instead of replacing the entire brief on every attempt.

Production-verified output

A real GPT Image 2 provider acceptance result.

Public brief summary: a controlled studio composition used to verify provider submission and image output handling; the complete acceptance prompt remains private.

Verified
2026-08-17T07:17:08.283Z
Public prompt hash
dd10c65895e9…
Acceptance settings
text-to-image workflow · 1:1 · 1K

Provider task succeeded with one image output during the documented production acceptance. This single production acceptance proves the connected task path, not universal quality, prompt accuracy, or repeatability for every image brief.

gpt image 2 production acceptance output in CreateForge
Public evidence asset from the production acceptance matrix; private IDs, full prompts, and temporary provider URLs are excluded.

What GPT Image 2 does inside CreateForge

GPT Image 2 supports both new image generation and reference-guided editing in CreateForge. Use text when the model can invent the scene, or add private references when identity, layout, product geometry, or an existing visual must influence the result.

The connected workflow exposes aspect ratio and resolution controls, validates the request on the server, and returns a short-lived quote before generation. Resolution is a production choice that can change cost, so early reviews and final delivery do not need the same setting.

After a successful provider task, eligible output is stored in the authenticated user’s private Library. Review exact text, identity, product details, and factual claims before publishing any generated image.

GPT Image 2 use cases

Choose the model for a specific deliverable.

01

Product campaign images

Build studio or lifestyle product compositions with deliberate camera, material, light, and copy space.

02

Reference-guided edits

Preserve identity, geometry, or layout while changing environment, material, wardrobe, or visual treatment.

03

Editorial and social assets

Create covers, portraits, poster imagery, and channel-specific frames from one structured creative brief.

Generation controls

A consistent form, tuned to the model.

Prompt

required

Describe the image to create with GPT Image 2.

Generation mode

required

Choose the input workflow before adding reference media.

Text to Image · Image to Image

Reference images

required

Accepts up to 16 image files.

Aspect ratio

required

Choose a supported output frame with visual ratio cards.

Auto · 1:1 · 3:2 · 2:3 · 4:3 · 3:4 · 5:4 · 4:5 · 16:9 · 9:16 · 2:1 · 1:2 · 3:1 · 1:3 · 21:9 · 9:21

Resolution

required

Your selected resolution is included in the credit price.

1K · 2K · 4K

Output count

required

Choose how many results to create. Each completed result is saved to your Library.

1 · 2 · 3 · 4

Prompt blueprint

  1. 01Open with the deliverable, subject, and non-negotiable constraints before adding style or finish.
  2. 02Explain what each reference controls and separate protected properties from requested changes.
  3. 03Select the final aspect ratio in the form and describe how the composition should occupy that frame.

Product hero

Create a premium studio hero of a cobalt ceramic speaker on pale stone, three-quarter camera, soft daylight from the left, realistic matte texture, warm-gray background, clean headline space on the right, no extra objects or text.

Defines the deliverable, subject, camera, light, material, and copy area in priority order.

Reference material edit

Use the supplied chair as the exact geometry and camera reference. Preserve frame, joinery, dimensions, and shadow direction. Change only the upholstery to woven deep-red wool with visible fiber detail, neutral gallery background, no added decoration.

Separates protected source properties from one bounded material change.

Explore prompt examples Open the image tool

Related guides

Learn GPT Image 2 by task.

View all guides

Field guide

Best AI image model: how to choose for real creative work

Find the best AI image model for a real deliverable by comparing reference control, typography, resolution, iteration cost, repeatability, and saved results.

Read guide

Model guide

How to use GPT Image 2 in CreateForge

A production-focused walkthrough for creating and editing images with GPT Image 2, including references, resolution choices, prompt structure, quoting, and result review.

Read guide

Prompt library

GPT Image 2 prompt examples for practical image work

Build a GPT Image 2 prompt with 12 practical examples for products, editorial images, typography, portraits, reference edits, and campaign assets.

Read guide

GPT Image 2 FAQ

Know the real product boundary.

Yes. Upload private references, explain what each source controls, list the properties that must remain, and describe the allowed transformation.

The connected CreateForge controls show the current supported resolution tiers. Select the delivery setting and review the protected quote before submitting.

No. GPT Image 2, GPT-Image 2, and GPT-Image 2.0 are consolidated on this canonical model page rather than split into duplicate URLs.

Eligible successful output is persisted to the authenticated user’s private CreateForge Library after provider and storage reconciliation.