GPT Image 2 API for Production Image Workflows
Generate or revise images through one API route with prompt, aspect-ratio, resolution, quality, and optional reference-image controls. GoEnhance provides independent access to OpenAI's GPT Image 2 model for asynchronous developer workflows.
Create an API Key
Build Image Generation and Editing into Your Product
Prompt-Led Image Generation
Send a prompt of 1 to 5,000 characters and choose an aspect ratio, resolution, and quality level. The strongest fit is controlled asset generation where your application can validate inputs, store the returned task ID, and review outputs before release.
Reference-Guided Image Work
Attach up to four HTTPS reference images when a task needs visual guidance. Treat references as constraints to evaluate, not guarantees: compare subject fidelity, composition, and unwanted carryover across outputs before moving an asset downstream.
Validate the documented request fields before submission to reduce avoidable failures.
GPT Image 2 API Parameters
Model
Set model to gpt-image-2. This exact identifier is required by the endpoint.
Prompt
Provide a required text prompt between 1 and 5,000 characters.
Aspect Ratio
Choose auto or a documented ratio from 1:1 through 21:9.
Image Size
Select 1k, 2k, or 4k output resolution; 1k is the documented default.
Quality
Select low, medium, or high; low is the documented default and pricing varies by tier.
Reference Images
Optionally pass up to four HTTPS image objects in image_list.
GoEnhance bills at $0.02 per token. The documentation lists these exact per-image rates by quality and resolution.
GPT Image 2 API Pricing
Low Quality
1k: 0.39 token ($0.008); 2k: 0.78 token ($0.016); 4k: 1.17 tokens ($0.023).
Medium Quality
1k: 3.45 tokens ($0.069); 2k: 6.9 tokens ($0.138); 4k: 10.35 tokens ($0.207).
High Quality
1k: 13.72 tokens ($0.274); 2k: 27.44 tokens ($0.549); 4k: 41.16 tokens ($0.823).
Integrate the GPT Image 2 API in Three Steps
1
Submit the Request
POST to /api/v1/images/generations with bearer authorization, model, prompt, and your chosen output controls.
2
Persist the Task ID
Read img_uuid from the response and store it with your own request record so retries and audits remain traceable.
3
Handle Completion
Poll GET /api/v1/jobs/detail or provide a public HTTPS callback URL, then process success or failure as a terminal state.
Use GPT Image 2 where structured requests and reviewable outputs matter more than one-off manual generation.
Where the API Fits Best
Product Visuals
Generate concept-ready product scenes for a review and selection pipeline.
Campaign Variants
Create controlled visual directions for ads while keeping approval in the loop.
Editorial Artwork
Turn written briefs into illustration candidates for articles and newsletters.
Game Concepts
Explore environments, props, and original character directions from structured prompts.
Reference Iteration
Use visual inputs to guide revisions, then inspect fidelity and accidental carryover.
Creative Automation
Queue image tasks from internal tools, content systems, or asset-production services.
Answers to common integration, pricing, and workflow questions.