Gemini Omni Flash resolution: "4K" or "4k" on Sume

Either spelling works. For Gemini Omni Flash 1.1 on Sume, the catalog token is 4K, and the request validator also accepts 4k as an alias. Before the job reaches the provider, the payload builder translates the catalog token to the lowercase 4k value the upstream expects, so the casing you type is never the casing that travels.

The practical advice is to send exactly what GET /v1/videos/models lists and treat the alias as a convenience.

What Google says 4K means

Google's Omni API page lists resolutions of 360p, 720p (the default), 1080p and 4K, and marks 1080p and 4K as upscaled (read 2026-10-04). Its launch post says the same about 1080p and 4K being upscaled, and says 360p drafts are faster and cheaper than 720p (read 2026-10-04).

So 4K on Omni is an output size, not a native render size. If you need native detail, 720p is the base.

The Sume catalog row

Sume lists four resolutions for the model: 360p, 720p, 1080p and 4K, with 720p as the default. Durations run 3-10 s, aspect ratios are 16:9 and 9:16 only, and audio is always on.

Cost of choosing 4K

Because 4K is an upscale, it is the most expensive way to get the same underlying render. At 8 s it bills $3.00, three times the 720p price of $1.00. The amount is reserved on submit and refunded if the job fails.

Use the 360p tier to check the shot first; the stored note on Omni 360p draft cost has the numbers.

List the resolutions from the API

This script prints the resolutions the live catalog reports for any Omni row.

import os
import requests

resp = requests.get(
    "https://api.sume.com/v1/videos/models",
    headers={"Authorization": f"Bearer {os.environ['SUME_API_KEY']}"},
    timeout=30,
)
resp.raise_for_status()
for m in resp.json()["data"]:
    if "omni" in m["id"]:
        print(m["id"], m["supported_resolutions"])
        print(m["pricing_skus"])