One prompt, three ratios: a 1:1, 16:9 and 9:16 set priced by model

Every one of the 19 models in Sume's image catalog lists 1:1, 16:9 and 9:16, so you can ask any of them for a square, a landscape and a portrait version of the same idea. A three-image set costs $0.075 on Qwen Image, about $0.137 on GPT Image 2.5 at high quality, and $0.5625 on Nano Banana Pro. The Image API docs list the aspect-ratio descriptors per model.

A set is three calls, one per ratio. The ratio is a request field, so the prompt stays identical and only aspect_ratio changes.

What does one set cost?

Flat models cost the same at any ratio. GPT Image 2.5 does not, because it is billed on output tokens: a landscape or portrait frame at 1280x720 is smaller than a 1024x1024 square. The GPT row below assumes explicit image_size pixels for each ratio.

What can differ across the three outputs?

The pixel dimensions are the model's native size for each ratio, not a fixed 1920x1080 or 1080x1920. If a platform needs exact pixels, plan a resize or crop step after generation. Composition also shifts: a prompt that works for a wide frame may crop a subject out of a tall one, so put the subject in the middle third and say so.

How should I run it?

Draft the set on a cheap model first, check the three crops, then rerun the keepers on the expensive one. Read usage.cost from each response so your budget tracks what Sume actually billed. Prices here are list x 1.25 in dollars per image before whole-cent rounding; the catalog's billable formula reads "list x 1.25, ceil usd cents", so confirm the first charge in usage.cost and budget from the invoice, not from the sum.