FLUX 3 video 20-second clips: what Sume's video API offers
FLUX 3 Video makes up to 20 s clips at 1920x1088. Sume does not list it, but seedance-2.5 and wan-3.0 accept 20 s requests. Durations by model.

BFL's FLUX 3 Video makes clips up to 20 seconds, and Sume does not list it. If a 20-second single request is what you need, Sume's video docs say seedance-2.5 accepts 4 to 30 seconds and wan-3.0 accepts 2 to 30 seconds, so 20 seconds fits either.
BFL's side comes from its release notes, Sume's from the Video generation docs; both were read 2026-09-29.
What does FLUX 3 Video offer?
BFL lists text-to-video, image-to-video, video continuation and draft enhance, clips up to 20 seconds at 1920x1088 and 24 fps, with synchronized multilingual dialogue. The endpoint is POST /v1/flux-3-video on BFL's API.
Which Sume video models take a 20-second request?
Durations differ by model, so read supported_durations from GET /v1/videos/models before you submit.
| Model id | Accepted length | Resolutions |
|---|---|---|
seedance-2.5 | 4 to 30 s | 480p, 720p, 1080p |
wan-3.0 | 2 to 30 s | not stated on the page |
minimax-h3 | 5 to 15 s | native 480p/768p |
gemini-omni-flash-1.1 | 3 to 10 s | 360p, 720p, 1080p, 4K |
| Every other catalog model | up to 15 s | see the catalog |
What does the request look like?
Video generation is asynchronous: submit, then poll the job. Set duration to the length in seconds.
curl -X POST https://api.sume.com/v1/videos \
-H "Authorization: Bearer $SUME_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model": "seedance-2.5", "prompt": "a slow dolly-in on a rain-lit street", "duration": 20}'Is this a like-for-like swap?
No. Each model has its own look, audio behavior and reference inputs, and the docs do not map one onto another. Treat a 20-second seedance-2.5 clip as its own result and judge it on your prompt.
What about resolution and frame size?
BFL states 1920x1088 for FLUX 3 Video at 20 seconds. On Sume, resolution comes from each model's supported_resolutions and aspect ratio from supported_aspect_ratios, both returned by the catalog. The docs give seedance-2.5 at 480p, 720p and 1080p, so a 1080p request at 20 seconds sits inside the documented limits, but sizes are named as tiers such as 1080p rather than as pixel dimensions.
What if I need a longer clip than one request allows?
Sume's longest documented single request is 30 seconds. Past that, generate more than one clip and join them; extend an AI video past 30 seconds and make a one-minute AI video walk through the steps.
BFL lists video continuation as a FLUX 3 Video mode of its own, which is a separate feature on BFL's API and not something Sume's endpoint exposes under that name.
How do I test a 20-second request cheaply?
Start with the lowest resolution the model lists, since seedance-2.5 accepts 480p, and review the motion and pacing there before you pay for a higher tier. Read supported_durations first, submit the job, poll it until it completes, then download the result. If the clip is close but not right, adjust the prompt and repeat at the same low tier so the comparison stays like for like.
Sources
Related posts
More in Models
- Gap or click when joining audio files: use wav, not mp3
Sume's Timeline audio joins up to 20 Sume-hosted parts with no silence at the seams. Its docs say to keep wav, because mp3 adds padding at every edge.
- Gemini Live Avatar vs a video avatar API: live or rendered?
Gemini 3.8 Live Avatar streams a lip-synced face in a conversation. Sume's avatar video renders a finished clip from a script. Which job needs which.
- Gemini Live Avatar: 97 languages and SynthID, explained
Google says Gemini 3.8 Live Avatar covers 97 languages and adds SynthID to audio and video. What each claim covers, and what it says about other avatar tools.
- Gemini Omni 1.1 Flash 360p drafts: what a cheap preview costs
Gemini Omni Flash 1.1 can render 360p previews. On Sume a 360p second is billed at 30% of a 720p second, so drafts cost less than finals.
Written by Sume