Luma uni-1 output_format png or jpeg vs Sume's per-model enum

Luma uni-1 accepts png or jpeg for output_format, and when you omit it the model picks one from the prompt. On Sume, output_format is a per-model enum drawn from png, jpeg, webp or svg, so set it explicitly and read the model's list.

Luma facts are from its Models page; Sume facts are from Image models, read 2026-10-01.

What does Luma offer for images?

What does Sume accept for output_format?

The Image API doc lists png, jpeg, webp, or svg as the field's values, but a model only accepts the values its catalog descriptors list. In the catalog code, Recraft V4 is webp only, while the edit-capable models list png, jpeg and webp.

output_compression (0-100 for webp or jpeg) is in the schema but marked not served in v1, so do not rely on it.

Why set the format explicitly?

On Luma, leaving it out means the format can change with the prompt, which breaks a pipeline that expects one extension. The Sume docs do not state a default, so the same advice holds: send the value you want and let a mismatch fail as a 400 rather than surprise you.

curl https://api.sume.com/v1/images \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"openai/gpt-image-2","prompt":"a product photo","output_format":"jpeg"}' 

How do I find the formats for a model?

Read its descriptors from GET /v1/images/models/{model_id}/endpoints and branch on them. The post on GPT Image 2.5 output formats shows the same check for one model.