3:2 and 2:3 image models on Sume: photo ratios and prices

Fourteen of the 19 image models in the Sume catalog list 3:2 and 2:3 (read 2026-10-03): Higgsfield Soul, Nano Banana 2 and Pro, the three Seedream rows, Grok Imagine, Qwen Image and Qwen Image Max, Flux 2 Pro and Flex, Recraft V4 and both Ideogram rows. The cheapest are Qwen Image and Grok Imagine at $0.025 per image. GPT Image 2.5, GPT Image 2 and Imagen 4 do not list them, so ask them for 4:3 or 4:5 and crop.

Who lists 3:2 and 2:3

The list comes from the catalog descriptors and the per-image endpoint prices (list x 1.25), read 2026-10-03. Higgsfield Soul also lists both, at an estimated $0.005 per image.

Why 3:2 matters

3:2 is the proportion of a common camera sensor, which makes it a natural fit for photo-style output that you may later print or crop. It is wider than 4:3 and less wide than 16:9, so it also suits blog headers and product-in-context shots.

Models that do not list it

GPT Image 2.5 lists 1:1, 16:9, 9:16, 4:3, 3:4, 5:4, 9:8, 4:5 and auto. If you need a 3:2 look from it, ask for 4:3 and crop the sides, or use auto. Check the returned size before cropping.

A request

Qwen Image at 3:2, four takes, costs 4 x $0.025 = $0.10.

curl -s https://api.sume.com/v1/images \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "qwen/qwen-image", "aspect_ratio": "3:2", "n": 4, "prompt": "Documentary-style photo of a baker dusting flour over a wooden counter, window light, natural colors"}' 

Text in photos

OpenAI's image guide (read 2026-10-03) says its models can still struggle with precise text placement and clarity. Keep signs and labels out of the prompt, or add them in post.