Skip to main content
gpt-image-2-web is the lower fixed-price reverse-image route for text-to-image and reference-image generation. The current price is USD 0.03 per request; check the model marketplace before relying on that value. This is a YingTu model ID in the GPT Image 2 family. OpenAI’s official model page publishes the gpt-image-2 alias and the dated snapshot gpt-image-2-2026-04-21; a direct OpenAI migration must change the model value as well as the base URL.

Where this route fits

Use Web for product mockups, editorial images, social assets, background changes, and queued image jobs when the three documented sizes are sufficient. The route supports:
  • synchronous text-to-image generation;
  • synchronous one-source-image editing;
  • unified task-ID generation and editing with 1–16 reference images;
  • 1024x1024, 2048x2048, and 3840x2160 output;
  • one Base64 PNG result.

How quality behaves

The route forwards low, medium, high, or auto. The selected upstream determines how strongly that value affects the result, so accepted syntax is not a cross-provider quality guarantee. Choose gpt-image-2-vip when the application needs the route whose four values have been exercised across the complete documented workflow.

Connection strategy

Use synchronous Generate or single-image Edit when the caller can wait for a complete response. For asynchronous or multi-image work, submit /v1/images/tasks with action=generate or action=edit, add Idempotency-Key, persist task_id, and poll with a deadline.

Review Web parameters

Confirm sizes, fields, response shape, and omissions.

Run Web examples

Generate, queue, edit, and save a PNG.
Last modified on September 3, 2026