A 4K image request returns 202: poll the job and fetch the result

POST /v1/images is synchronous by default: it waits up to 30 seconds and returns 200 with data[].url images. When generation is slower, Sume returns 202 and a job envelope instead of an error. The Sume docs name the slow cases: 4K, high quality and large n. They also say to examine the status code, not the body shape.

The two shapes

Python that handles both

Poll GET /v1/jobs/{id}/status until it reports terminal, then read GET /v1/jobs/{id}/result. The status endpoint's sume_status is completed on success, and the result endpoint answers 409 job_not_completed for any other state.

import os, time, requests
base = "https://api.sume.com"
headers = {"Authorization": "Bearer " + os.environ["SUME_API_KEY"]}
body = {"model": "google/nano-banana-2", "prompt": "Mountain at dawn",
        "resolution": "4K", "aspect_ratio": "16:9"}
r = requests.post(base + "/v1/images", headers=headers, json=body, timeout=60)
if r.status_code == 200:
    print([d["url"] for d in r.json()["data"]])
elif r.status_code == 202:
    job_id = r.json()["data"]["job"]["id"]
    while True:
        s = requests.get(f"{base}/v1/jobs/{job_id}/status", headers=headers, timeout=30).json()["data"]
        if s.get("terminal"):
            break
        time.sleep(s.get("next_poll_after_seconds") or 3)
    if s.get("sume_status") == "completed":
        print(requests.get(f"{base}/v1/jobs/{job_id}/result", headers=headers, timeout=30).json())
else:
    print(r.status_code, r.text)

Alternatives

If you already know the call is slow, send mode: "async" and skip the wait. The job result keeps the platform's generic job shape for every product, so print it once to see where your images sit. Status and result bodies wrap their fields in data, which is why the loop reads ["data"].

How this was checked

Vendor facts come from the pages listed in the sources, read on 2026-10-05. Sume facts come from the Image API docs and the catalog code on main on the same date. Catalogs and limits change, so read the descriptors from GET /v1/images/models before you pin a number in production code.