Product ad with sound: Omni Flash $1.88 or Wan 3.0 $2.50 at 1080p
A 10-second 1080p product clip with audio costs $1.88 on Gemini Omni Flash 1.1 and $2.50 on Wan 3.0 on Sume; Wan reaches 30 s and takes audio references.

A 10-second 1080p product ad with sound costs $1.875 on gemini-omni-flash-1.1 and $2.50 on wan-3.0 on Sume, before the job is rounded up to the cent. Omni is the cheaper clip, but it stops at 10 seconds and takes no audio reference; Wan is the pick when the ad needs 11 to 30 seconds or your own audio.
Rates come from the Sume catalog rows (provider list times 1.25), read 2026-10-05 in the Video Router guide and Video Generation.
Price and limits at 1080p
Sume bills the provider list price times 1.25 and rounds the job up to the cent, so a one-second figure here is a rate, not a charge.
| Model id | Rate per second at 1080p | 10-second clip | Max length | Audio |
|---|---|---|---|---|
| gemini-omni-flash-1.1 | $0.1875 | $1.875 | 10 s | Native, always on |
| wan-3.0 | $0.25 | $2.50 | 30 s | Yes; audio references up to 5 |
| minimax-h3-max | $0.20 | $2.00 | 15 s | Native stereo, always on |
| kling-3 | $0.21 with audio | $2.10 | 15 s | generate_audio true |
Choosing between them
If the ad is 10 seconds or less and you want the lowest bill, take Omni. At 1080p it is also the only one of these with a native 4K option, which the rest of the catalog reaches only through H3 upscales.
If the cut runs 12 to 30 seconds, Wan takes it in one job and the cost is linear: 20 seconds at 1080p is $5.00. Kling and H3 Max stop at 15 seconds, so a 20-second ad on those means two clips and a join.
- Test the prompt at 480p first: Wan is $0.0625 per second there.
- Send
generate_audio: falseto Kling or Wan if you will lay a music bed yourself. - Omni and H3 Max have no audio toggle, so the sound is in the price.
Request
Pin the model and the 1080p tier explicitly; sume/auto would pick Omni 720p by default.
import os, time, requests
H = {"Authorization": f"Bearer {os.environ['SUME_API_KEY']}"}
payload = {
"model": "wan-3.0",
"prompt": "Product ad: a stainless water bottle on a trail rock at sunrise, slow push-in, light wind",
"duration": 10,
"resolution": "1080p",
"aspect_ratio": "9:16",
"generate_audio": True,
}
job = requests.post("https://api.sume.com/v1/videos", headers=H, json=payload).json()
while True:
time.sleep(30)
s = requests.get(job["polling_url"], headers=H).json()
if s["status"] in ("completed", "failed", "cancelled"):
break
print(s["status"], s.get("unsigned_urls"), s.get("usage"))Sources
Related posts
More in Pricing
- Sume queue_full: is money held for the job that was rejected?
A 429 queue_full means the workspace has no accepted-job capacity left. Sume releases or refunds the reservation for the failed admission. How to confirm it.
- Quiz audio: 100 questions at 140 characters, one job, sliced per line
A hundred 140-character quiz questions are 14,000 characters: $0.21 on MAI Flash, $0.31 on MAI-Voice-2.1 and $0.67 on Sume, which can slice one job by sentence.
- Quote a 30-second AI video clip: Seedance 2.5 per clip and per minute
How to quote a 30-second Seedance 2.5 clip on Sume: price at 480p, 720p and 1080p, a per-minute equivalent, a retake allowance and the rounding rule.
- Re-render your top old videos first under a fixed $3 budget
Rank clips from a retired video pipeline by views per dollar of Sume re-render cost, then fill a fixed budget greedily. Python, arithmetic shown.
Written by Sume