Vidu Q4 clip lengths 3 to 16 s vs every Sume video row's limits
Vidu's image-to-video page allows 3-16 seconds. Sume has no Vidu row; this table maps that range to each Sume row's minimum and maximum, and who reaches 16.

Vidu Q4 Preview's image-to-video endpoint accepts 3 to 16 seconds, and on Sume only Seedance 2.5 and Wan 3.0 can produce a 16-second text-or-image clip; Sume has no Vidu row at all. Vidu's parameter page lists duration as 3 - 16 with a default of 5 (Vidu API, read 2026-10-10).
So the question for a migration is not which model looks similar but which Sume rows keep your existing duration values valid. This post lines the ranges up so you can see where a Vidu job needs a different length.
What the Vidu page gives
The same page lists resolutions 540p, 720p, 1080p, 2K and 4K with a default of 720p, one starting image of up to 50 MB (PNG, JPEG, JPG or WEBP, as Base64 or an accessible URL), a 20 MB cap on the whole HTTP body, and an audio option: with audio enabled the output includes dialogue and sound effects, disabled gives a silent video. The page does not list prices or aspect ratios in the text we read, so we make no price claim for Vidu here.
A second launch-week point: this is a Preview. Treat any duration you hard-code as something that can change.
Sume's ranges
Sume publishes ranges per row in supported_durations on GET /v1/videos/models. The catalog in the repo gives the following ranges, and docs.sume.com says each limit differs by model (Video Generation). One docs line says every other model tops out at 15 s; Gemini Omni Flash 1.1 is 3 to 10 s, so that line is not a safe guide for it.
| Row | Min s | Max s | Covers Vidu's 16 s? |
|---|---|---|---|
| Vidu Q4 Preview (Vidu) | 3 | 16 | Yes |
| Seedance 2.5 | 4 | 30 | Yes |
| Wan 3.0 | 2 | 30 | Yes |
| Seedance 2.0, Fast, Mini | 4 | 15 | No |
| Kling Video v3 Pro | 4 | 15 | No |
| Grok Imagine Video 1.5 | 4 | 15 | No |
| MiniMax H3 and H3 Max | 5 | 15 | No |
| Gemini Omni Flash 1.1 | 3 | 10 | No |
| Veo 3.1 (three rows) | 4 | 8 (4, 6 or 8 only) | No |
Which edge cases bite
Three cases come up when you port a Vidu job:
- A 3-second Vidu clip: Wan 3.0 (min 2) and Omni Flash (min 3) accept it; Seedance, Kling, Grok and Veo start at 4, and MiniMax starts at 5.
- A 16-second clip: only Seedance 2.5 and Wan 3.0 accept it. Splitting into 8 + 8 on another row is a different edit, not the same shot.
- Veo takes 4, 6 or 8 only, so 5 or 7 seconds is rejected rather than rounded.
A guard that fails early
The simplest protection is to read the row's ranges at runtime and refuse a request before you spend anything. Sume also validates server-side, so a wrong duration returns a 400 before a reserve; the guard just gives you a clearer message in your own logs.
If you let Sume choose, sume/auto is limited to 3 to 10 second clips at 16:9 or 9:16 with a 720p, 8 second default, so a 16-second job must pin a row (Video Router).
Sources
Related posts
More in Comparisons
- Vidu Q4 has no text-to-video: Sume rows that take a plain prompt
Vidu's Q4 page lists only image-to-video and reference-to-video. Sume has no Vidu row, but several rows start from a text prompt alone. Compare limits.
- Vidu Q4 takes 15 reference images; Sume rows cap at 9 or 10
Vidu Q4 accepts 1 to 15 reference images. Sume has no Vidu row: most rows allow 9 images, Wan 3.0 and Omni allow 10. Caps for images, video and audio refs.
- Vidu Q4 voice references vs Sume's reference audio inputs
Vidu Q4 takes up to 3 voice references. Sume has no Vidu row; Seedance, Wan 3.0 and MiniMax H3 accept audio references, but voice cloning is app-only, not API.
- Voxtral Mini 4B Realtime Arabic: open weights vs a Sume STT job
Mistral's Arabic streaming model arrived Oct 8 as Apache 2.0 weights, not a service. For Arabic transcripts without hosting, Sume STT takes language_code ar.
Written by Sume