Shortest clip each Sume video row accepts, from 2 s Wan to 5 s H3
Minimum duration per Sume video row, from 2 seconds on Wan 3.0 to 5 on H3, with the billed price of that shortest clip at the cheapest resolution.

The shortest clip on Sume is 2 seconds on Wan 3.0; Gemini Omni Flash 1.1 and sume/auto start at 3, most other rows at 4, and the MiniMax and Recast rows at 5. That matters when a launch-week model such as Vidu Q4 Preview advertises a 3-second floor and you want to know which Sume rows can match it.
The ranges come from the catalog in the repo and from Video Generation; confirm live values in supported_durations.
The ladder
This table lists each row's minimum and the billed price of that minimum at the cheapest resolution, using the list rate in the repo times 1.25. Seedance rows are billed on video tokens, not on a flat per-second rate, so they appear without a computed price.
| Row | Min seconds | Cheapest resolution | Billed per second | Billed shortest clip |
|---|---|---|---|---|
| Wan 3.0 | 2 | 480p | $0.0625 | $0.125 |
| Gemini Omni Flash 1.1 | 3 | 360p | $0.0375 | $0.1125 |
| Grok Imagine Video 1.5 | 4 | 480p | $0.0125 | $0.05 |
| Veo 3.1 Lite, no audio | 4 | 720p | $0.0375 | $0.15 |
| Kling Video v3 Pro, no audio | 4 | 1080p | $0.14 | $0.56 |
| MiniMax H3 | 5 | 480p | $0.0625 | $0.3125 |
| MiniMax H3 Max | 5 | 480p | $0.0625 | $0.3125 |
| Seedance 2.5 / 2.0 / Fast / Mini | 4 | 480p | Token-billed | See estimate |
Rows with a different shape
Three rows do not fit a plain minimum. H3 Max Recast needs a source video of 5 to 30 seconds, and the duration you send is that source length rounded up. Genjutsu Motion Transfer accepts 4 to 30 seconds and is in the catalog only when its provider is configured. Veo 3.1 accepts only 4, 6 or 8, so its floor is also its first step.
Using the floor in practice
Short clips are cheap, but a short floor is not the same as a short best price. Two cases show why:
- A 2-second logo sting fits Wan 3.0 at $0.125 billed at 480p, or Omni at 3 seconds and trim to 2 in the timeline.
- A 4-second loop is cheapest on Grok at $0.05, but only from a still.
- A 5-second text-only B-roll can go to MiniMax H3 at $0.3125 at 480p, or to Wan at $0.3125 for the same 5 s at 480p, so the choice is about look, not price.
How to read it in code
Do not copy this table into your app. Read supported_durations from GET /v1/videos/models and compare it to your request before you submit. A request below the minimum fails validation before any balance is reserved, so it is a quick error, not a charge.
Sources
Related posts
More in Models
- Aspect ratio auto on Sume: 6 image models list it, 8 edit models don't
Which Sume image models accept aspect_ratio auto when you edit a photo, which edit models reject it with a 400, and how to pick a ratio in code instead.
- Sume image catalog on 2026-10-10: 19 ids from 0.5 to 26 cents
All 19 Sume image model ids sorted by base price, with how many take references, lists auto, offer a quality field, a resolution field, or a mask.
- Is Utopai X on Sume? It is built on MiniMax H3, which Sume lists
Utopai X runs inside Utopai's PAI platform and is post-trained on MiniMax H3. Sume does not list Utopai X; it lists minimax-h3 and minimax-h3-max.
- Veo 3.1 4K needs an 8-second clip; Sume's Veo rows stop at 720p
Google offers Veo 3.1 1080p and 4K only at 8 seconds. Sume's Veo rows run 720p at 4, 6 or 8 seconds. See the gap and which Sume row reaches 4K.
Written by Sume