GPT Image 2 Official

gpt-image-2-official is the official GPT Image 2 channel variant for image generation and editing with an explicit quality tier.

Its key difference from gpt-image-2 is the quality field; do not send that field to the standard model.

Quick Facts

Item	Value
Model ID	`gpt-image-2-official`
Type	Image
Region	Global
Best For	Official image generation with controllable quality
Priority	Use when `quality` is required

Recommended Request Shape

Submit to POST /v1/generations and choose a quality tier for the target cost and result.

{
  "model": "gpt-image-2-official",
  "prompt": "A premium skincare poster on warm marble, luxury editorial style",
  "quality": "high",
  "resolution": "2K",
  "n": 1
}

Field Reference

Field	Type	Required	Default	Description
`model`	`string`	Yes	-	Must be `gpt-image-2-official`
`prompt`	`string`	Yes	-	Image prompt
`reference_images`	`string[]`	No	-	Reference image URL list
`quality`	`string`	No	`medium`	`low` is speed/cost-first, `medium` is balanced, and `high` is quality-first; the tier affects cost
`resolution`	`string`	No	`1K`	Output resolution
`size`	`string`	No	-	Upstream size field
`n`	`integer`	No	`1`	Image count

Supported Enums

Field	Values
`quality`	`low` (speed/cost-first), `medium` (default balance), `high` (quality-first, higher cost)
`resolution`	`1K` `2K` `4K`

Integration Notes

Start with quality=medium and resolution=1K during integration.
quality is official-model-only; remove it when switching to gpt-image-2.
Use n for multiple images; billing follows the actual image count.

Smallest Example

{"model":"gpt-image-2-official","prompt":"A premium skincare poster on warm marble"}