27-second AI video API: Wan 3.0 price at each resolution
A 27-second clip fits one job on Sume only with Wan 3.0 or Seedance 2.5. Exact Wan 3.0 prices at 480p, 720p and 1080p, plus the request body.

Two models on Sume accept a 27-second request: Wan 3.0 (wan-3.0, 2 to 30 seconds) and Seedance 2.5 (seedance-2.5, 4 to 30 seconds). Every other model in the video catalog tops out at 15 seconds, so a 27-second job on any other id is outside that model's listed range. For Wan 3.0 the price is $1.69 at 480p, $3.38 at 720p and $6.75 at 1080p.
Which models accept 27 seconds
The video models listing (GET /v1/videos/models) reports supported_durations per model. The Sume docs state the limits in one sentence: seedance-2.5 accepts 4 to 30 seconds at 480p, 720p and 1080p, wan-3.0 accepts 2 to 30 seconds, and every other catalog model except the source-video models tops out at 15 seconds.
- Wan 3.0: 2 to 30 s, 480p, 720p, 1080p, aspect ratios auto, adaptive, 16:9, 4:3, 1:1, 3:4 and 9:16.
- Seedance 2.5: 4 to 30 s, 480p, 720p, 1080p, with 21:9 added to the same list.
- Genjutsu and H3 Max Recast also reach 30 s, but they transform a source video you supply, so their length follows that clip rather than a duration you choose.
What Wan 3.0 costs for 27 seconds
Sume bills the provider list price multiplied by 1.25 and rounded up to the cent. The balance is reserved when you submit, so you can read the figure below before the job starts. 27 s sits three seconds under the 30 s ceiling.
| Resolution | Provider list per second | Provider list for 27 s | Sume price (list x 1.25, rounded up to the cent) |
|---|---|---|---|
| 480p | $0.05 | $1.35 | $1.69 |
| 720p | $0.10 | $2.70 | $3.38 |
| 1080p | $0.20 | $5.40 | $6.75 |
Submit it
The request is the OpenRouter-compatible POST /v1/videos. Send an Idempotency-Key so a retry after a timeout does not create a second billable job.
curl -X POST "https://api.sume.com/v1/videos" \
-H "Authorization: Bearer $SUME_API_KEY" \
-H "Content-Type: application/json" \
-H "Idempotency-Key: wan-3.0-27s-001" \
-d '{
"model": "wan-3.0",
"prompt": "A slow dolly shot through a greenhouse at dawn",
"duration": 27,
"resolution": "480p",
"aspect_ratio": "16:9"
}'Check the limit before you pay
Do not hard-code the 30 s ceiling. Read supported_durations from GET /v1/videos/models at runtime and fall back to a shorter clip if the id you want is absent. Check the envelope first so a request outside it is not submitted at all.
Where to go next
If your plan needs more than 30 seconds, split it into scenes and join the clips with the Timeline endpoint ($0.10 per output minute in the Sume docs). The related posts below cover the neighbouring lengths one by one.
Sources
Related posts
More in Models
- 29-second AI video clip: Wan 3.0 or Seedance 2.5 on Sume
Need 29 seconds of AI video in one render? Two Sume models take it. Wan 3.0 costs $1.82 at 480p, $3.63 at 720p and $7.25 at 1080p.
- 31-second AI video API: why Sume needs two jobs past 30 seconds
No Sume video model takes duration 31. Wan 3.0 and Seedance 2.5 stop at 30 seconds; here is the two-job route, what it costs and how to join the clips.
- 7-second AI video API: which Sume ids take 7 s and the price
Veo documents 4, 6 and 8 second clips, so 7 is not an option there. All twelve Sume ids accept 7 s; Omni Flash 1.1 at 360p is cheapest.
- 9-second AI video API: Omni takes 9 s, Veo does not
A 9 second clip fits Gemini Omni Flash 1.1 (3-10 s) and most Sume ids, but not Veo 3.1, which tops out at 8 s. Billed prices for 9 s inside.
Written by Sume