Image model
GPT Image 2
Premium image generation and editing with GPT Image 2: text-in-image, 1K–4K quality tiers, and reference-still edits inside the CircleAIDance composer.
Internal CircleAIDance sample · generated with GPT Image 2 · prompt: “Luxury perfume bottle on a dark stone surface, soft smoke, dramatic studio light, shallow depth of field, premium product photography”
Specifications
- Scene
- Text to Image · Image to Image
- Aspect ratios
- 1:1 · 16:9 · 9:16 · 3:2 · 4:3 · 3:4 · 21:9
- Resolutions
- 1K · 2K · 4K
- Images per run
- NaN–NaN
- Credits per image
- 6–10
What GPT Image 2 does not do
- Video output
- It generates still images only; animation needs an image-to-video model like Hailuo 02 or Kling 3.0.
- Native audio
- No audio track ships with any output.
- Free unlimited runs
- Every run bills flat per image by resolution; there is no unlimited tier.
GPT Image 2
GPT Image 2 is the premium model in the CircleAIDance composer, for frames that need fine typography, dense detail, or print-grade resolution. It handles text-to-image and image-to-image in one place, with 1K, 2K, and 4K quality tiers on supported formats.
Costs scale with the quality tier you pick, and the composer always shows the exact per-image credits before you generate — the same figures the pricing page uses.