Image quality default on Sume: low on Image 1.0, high on GPT 2.5

On Image 1.0, omitted quality means low. On POST /v1/images with ChatGPT Image 2.5, omitted quality defaults to high. The same omission gives different results on the two routes, so set quality explicitly.

From Sume's Image 1.0 and Image API docs, read 2026-09-30.

What are the values?

Image 1.0 accepts low (default), medium, high, and advises escalating for finals, dense text or packaging. The Image API lists auto, low, medium, high, xhigh or max, catalog-gated per model.

What does auto do?

For ChatGPT Image 2.5 the docs say auto quality reserves max cost up front. If you want a predictable reserve, pass an explicit level instead of auto.

How do I pin it?

Set the field on every request:

curl -X POST https://api.sume.com/v1/images \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "openai/gpt-image-2.5",
    "prompt": "Shampoo bottle label with the words Fresh Start",
    "quality": "medium"
  }'

When should I raise it?

Image 1.0's docs say to escalate quality for finals, dense text or packaging. Draft at a lower level, then re-run the chosen prompt at a higher one; a model only accepts the levels its descriptors list.

How do I check this myself?

Because defaults differ, a prompt that looked fine on Image 1.0 may be billed differently when moved to /v1/images without a quality value. Compare the endpoint's pricing lines for the levels you plan to use before you switch. The linked docs pages and the catalog endpoint show the current values, and this post reflects them as of 2026-09-30.