6 product photos as video references: Sume models vs Vidu Q4's 15

If you have six product photos and want one video that keeps the product consistent, the models that take reference images on Sume are Seedance 2.5, Seedance 2.0, Seedance 2.0 Fast, Seedance 2.0 Mini, Wan 3.0, MiniMax H3, MiniMax H3 Max and Gemini Omni Flash 1.1. Vidu Q4 Preview, launched 2026-10-07 with up to 15 image references, is not in Sume's video catalog (read 2026-10-08).

What Vidu announced

The Vidu Q4 Preview release (read 2026-10-08) says the model takes up to 15 image references and up to 3 audio references, outputs 540p, 720p, 1080p, 2K or 4K at 10-bit color, and starts at $0.014 per second on Vidu's web app and API. That is Vidu's own price, not a Sume price. At that rate a 5-second clip is 5 x $0.014 = $0.07 on Vidu. Sume does not list a Vidu model, so there is nothing to compare it with inside Sume.

Models on Sume that accept reference images

The Sume catalog marks reference images per model. Send them as input_references on POST /v1/videos, or as reference image URLs on the Video Router. A model only accepts a reference type if its supported_input_references lists it, so read that field before you build a six-photo request.

Five seconds with six photos

A reference-to-video call treats the images as visual guidance, not exact frames. If you need the first frame to be your packshot, use frame_images with frame_type: first_frame; when you send both fields, frame_images decides the mode and Sume runs image-to-video. The 5-second price depends on the model and resolution you pin.

A practical way to start

Pick the hero angle for frame_images and pass the other photos as references. Test at the cheapest resolution first, as in the Seedance Mini row above. All reference URLs must be public HTTPS. The Wan 3.0 limits (10 images, 5 video clips, 5 audio clips) are the only per-type counts Sume documents for these models, so for the others, read the catalog entry before you send more than a few images.