Ideogram Z-Image resolution rules vs Sume image_size

Ideogram's Z-Image endpoint takes resolution as WIDTHxHEIGHT, each side a positive multiple of 16, total area at most 1,048,576 pixels (1024×1024). Sume's images API uses normalized tiers (512, 1K, 2K, 4K) with aspect_ratio; custom pixels exist only as image_size on ChatGPT Image 2.5, where both edges must be multiples of 16.

Ideogram facts are from its API reference; Sume facts from Image generation, read 2026-10-01.

What are Ideogram Z-Image's limits?

POST /v2/image/generate/z-image takes a JSON body with a required prompt. resolution is optional; when omitted the default is 1024×1024. Other fields include num_inference_steps, seed, num_images (default 1), async and webhook_url.

How do the size rules compare?

Which Sume field do I send?

For most models send resolution and aspect_ratio; the docs say size is shorthand for a tier and must not carry custom pixels. Only ChatGPT Image 2.5 accepts custom pixels through image_size. The docs also say a model accepts only the values its catalog descriptors list, so read supported_parameters before pinning one. See size vs image_size for the details.

What about quality and unlisted fields?

quality accepts auto, low, medium, high, xhigh or max, gated by the catalog. A request that sets a parameter the selected model does not list is rejected, so a Z-Image-style num_inference_steps will not be silently ignored.