Recraft Studio batch of 1, 2, 4, 6 or 8 images vs Sume's n range

Recraft's changelog says you can choose a batch size of 1, 2, 4, 6 or 8 images before you generate in Studio. The Sume image API takes n from 1 to 10 at request level, but each model has its own ceiling published as a range descriptor, so the right number comes from the catalog. A fixed menu of batch sizes is a Studio feature; on Sume any integer inside the model's range is valid.

What the vendor says

From the Recraft changelog.

What Sume says

Per the Image API docs, you can request up to 10 images per call with n, but per-model ceilings are lower, so read the n range descriptor from the catalog. Billing is cost_usd times n, because endpoint pricing lines are what your wallet is charged.

Read the ceiling first

This script prints the n range for every model, so you can choose a batch size without a failed request.

import os, requests

r = requests.get(
    "https://api.sume.com/v1/images/models",
    headers={"Authorization": f"Bearer {os.environ['SUME_API_KEY']}"},
    timeout=30,
)
for m in r.json()["data"]:
    n = m.get("supported_parameters", {}).get("n")
    if n and n.get("type") == "range":
        print(m["id"], n["min"], n["max"])

Large batches

A big n is one of the slow configurations the docs name, so expect a 202 job envelope rather than an inline 200. Check the status code, not the body shape, and read images from the job result when it is 202. Failed generations are not billed.