Image model
Z-Image
Generate portraits, product shots, and social crops with Z-Image in the CircleAIDance composer. Flat per-image credits, five aspect ratios, up to four images per run.
Internal CircleAIDance sample · generated with Z-Image · prompt: “Editorial portrait of a woman looking over her shoulder, soft window light, dark tailored coat, 35mm film grain, detailed skin and hair”
Specifications
- Scene
- Text to Image
- Aspect ratios
- 1:1 · 4:3 · 3:4 · 16:9 · 9:16
- Resolutions
- standard
- Images per run
- NaN–NaN
- Credits per image
- 1
What Z-Image does not do
- Image-to-image edits
- Z-Image is text-to-image only; it does not restyle an uploaded photo. Use GPT Image 2 or Nano Banana 2.1 for image-to-image.
- Resolutions above its native size
- There is no 2K/4K toggle; for large print or heavy crops use GPT Image 2 2K instead.
- Native audio
- Z-Image outputs still images with no audio track.
Z-Image
Z-Image is the entry model in the CircleAIDance composer: a fast text-to-image model for drafts, portraits, and social formats. One prompt covers five aspect ratios, and a single run can return up to four variations so you can pick the strongest frame instead of rerolling a single image.
Every Z-Image run shows its exact credit cost before you generate, and finished images land in your History and Assets automatically — ready to download, or to re-edit with a different model later.