Ideogram 8K output resolution API vs Sume 512 to 4K tiers

Ideogram 4.0 does accept an 8K output: its output_resolution enum is 1k, 2k, 4k or 8k. Sume's image route has no 8K value. Its resolution field is a normalized tier of 512, 1K, 2K or 4K, so 4K is the ceiling there.

Ideogram facts are from its Ideogram 4.0 generate page; Sume facts are from Image models, both read 2026-10-01.

What does Ideogram's output_resolution actually mean?

The page says each tier is a total pixel budget equal to a square of the named size. Its example: 8k delivers at most 8192x8192 pixels in total. Wide and tall aspect ratios keep the same budget, so one side may exceed the named size. It also says tiers above 2k are produced by upscaling after generation, and the default is 1k.

How do the two schemes compare?

What does Sume do with a value it does not list?

Sume rejects it. The docs say a request that sets a parameter the selected model does not list is rejected with 400 unsupported_parameter rather than silently dropped. They also say a model only accepts the values its catalog descriptors list, so read supported_parameters before pinning a tier or ratio.

curl https://api.sume.com/v1/images \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"bytedance-seed/seedream-4.5","prompt":"a landscape photo","resolution":"4K","aspect_ratio":"16:9"}'

Will a 4K request come back inline?

Not always. POST /v1/images blocks for up to 30 seconds and returns 200 with the image response. The docs name 4K among the slow configurations most likely to degrade to 202, which is the standard job envelope; poll GET /v1/jobs/{id}/status then fetch GET /v1/jobs/{id}/result. Check the status code, not the body shape.

What should I do if I need more than 4K?

Read the per-model capability descriptors first; they are the source of truth for what a model accepts. For the 4K tier and the generate-or-upscale choice, see AI image 4K: generate or upscale. Do not send an 8k string to Sume expecting it to be mapped down.