19-second AI video with references: Wan 3.0 limits on Sume

You can send references with a 19 second Wan 3.0 request on Sume: up to 10 images, up to 5 videos and up to 5 audio clips. The catch is that the reference videos may total at most 15 s and the audio clips at most 15 s, so the output can be longer than the references.

Which models accept 19 seconds

Wan also requires reference videos to be at least 16 fps. Seedance 2.5 accepts references too, and it is the other generator that reaches 19 s.

Reference limits are separate from clip length

Output length and reference length are independent. A 19 s Wan clip can be guided by 15 s of reference video; it cannot be guided by 20 s.

The wire format depends on the endpoint. POST /v1/videos takes input_references and frame_images, with frame_images winning when both are set. The legacy POST /v1/video-router/generate takes flat fields such as image_url, end_image_url and video_url.

• Images: at most 10 per request.
• Videos: at most 5, at most 15 s combined, at least 16 fps.
• Audio: at most 5, at most 15 s combined.

What a 19 second request costs

Sume bills the provider list price times 1.25, rounded up to the next cent. The amount is reserved when you submit and refunded if the job fails, so a rejected or failed clip does not cost you. The /v1/videos/models descriptor lists the billable rate in pricing_skus, which is the field to trust over any number in a blog post.

A 19 s Wan clip bills $1.19 at 480p, $2.38 at 720p and $4.75 at 1080p. Check pricing_skus for any reference-related charge on your chosen resolution.

Check 19 seconds against the live catalog

Duration support lives in the catalog, not in this page. Each entry on GET /v1/videos/models carries supported_durations, which is every whole second from the model minimum to its maximum, so one membership test answers the question. This script lists every model that accepts 19 s, with its resolutions and billable rates.

Set SUME_API_KEY first. See the Video Generation docs for the request and polling lifecycle.

import os
import requests

N = 19
resp = requests.get(
    "https://api.sume.com/v1/videos/models",
    headers={"Authorization": f"Bearer {os.environ['SUME_API_KEY']}"},
    timeout=30,
)
resp.raise_for_status()
for m in resp.json()["data"]:
    if N in m["supported_durations"]:
        print(m["id"], m["supported_resolutions"], m["pricing_skus"])