Wan 3.0 at 30 fps and 2 to 30 s: the Sume model id

Alibaba's Model Studio documentation lists wan3.0-video and wan3.0-video-prime at 480P, 720P or 1080P, 2 to 30 seconds, 30 fps, MP4. On Sume the catalog id is wan-3.0, and it accepts 2 to 30 seconds. The Sume docs do not state a frame rate, so check the output with Video inspect if 30 fps matters to you.

What Model Studio lists

The Model Studio page says these models support text-to-video, first-frame and last-frame image-to-video, and reference generation, and quotes the output as "480P, 720P, 1080P. 2-30s. 30 fps, MP4".

Call it on Sume

Use the bare catalog id, with no provider prefix. Send duration as an integer number of seconds. To give a first frame, pass frame_images with frame_type: "first_frame"; add a second entry with last_frame for an end frame. If you pass both frame_images and input_references, frame_images wins and the request is image-to-video.

curl -X POST https://api.sume.com/v1/videos \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: wan-demo-001" \
  -d '{
    "model": "wan-3.0",
    "prompt": "A paper boat drifts down a rain gutter, camera follows at water level",
    "duration": 12,
    "frame_images": [{
      "type": "image_url",
      "image_url": {"url": "https://example.com/first-frame.png"},
      "frame_type": "first_frame"
    }]
  }'

Check what you actually got

Do not assume the Model Studio spec carries over unchanged to another host. Read GET /v1/videos/models for the live supported_durations, supported_resolutions and supported_aspect_ratios of wan-3.0. Then run Video inspect on a finished clip. It returns probe facts, so you can read the real duration and frame rate rather than trusting a table in a blog post.

When 30 seconds is the point

A single 30 second job avoids stitching seams, but it is one expensive attempt. If you only need 12 seconds, ask for 12. Billing on Sume is provider list times 1.25, and the catalog lists the price lines for the model.