Model Briefing

GPT Image 2

GPT Image 2 is an image model (not video). Use it to generate stills from a prompt or edit reference images, with quality set to auto, low, medium, or high.

Launch information

Last updated: 2026-08-16

Launched by: OpenAI

GPT Image 2 is OpenAI's next-generation image model, with text-to-image, image editing, quality tiers, and flexible aspect ratios.


At a glance

Speed: Medium

Quality tier: Premium

Audio support: None

Cost tier: Budget

Duration range: N/A (image model)

Resolution range: Flexible presets up to 4K


Best for

  • - Text-to-image with quality controls
  • - Image-to-image editing
  • - Detailed typography and product stills

Technical profile

Modes: Text-to-image and image-to-image

Quality: auto, low, medium, high

Aspect ratios: 1:1, 4:3, 16:9, 3:4, 9:16, auto

Reference images: Up to 12

Outputs: Up to 4 images per run

GPT Image 2 — AI Video Model | Speed, Quality & Examples | Spectoria