Vidu Q2 Pro Fast image-to-video 1080p vs Sume first frame

QwenCloud lists vidu/viduq2-pro-fast_img2video as an image-to-video model that supports 720P and 1080P. Sume's catalog has no Vidu id, so on Sume you send the image as image_url (the first frame) to a catalog model that offers 1080p, such as seedance-2.5.

The Vidu facts come from the QwenCloud model changelog; the Sume facts from the Video Router docs and video models guide, read 2026-10-01.

What does the Vidu entry actually say?

The changelog describes it as a Vidu image-to-video model: pass in an image and a text prompt to generate video using the image as the initial frame, with 720P and 1080P resolution. It says nothing in that entry about duration or price, so this post does not state either.

How do I pass a first frame on Sume?

On the Video Router shape, image-to-video is prompt plus image_url, and the image is the first frame. Add end_image_url to also pin the last frame. Use public HTTPS media URLs only.

On the OpenRouter-shaped /v1/videos body, the same job uses frame_images, where each entry carries a frame_type of first_frame or last_frame. If both frame_images and input_references are sent, frame_images wins and the request is image-to-video.

Which Sume ids reach 1080p?

From the Video Router table, these list 1080p. Limits differ per model, so read capabilities from GET /v1/video-router/models instead of assuming one envelope.

What should I do?

Keep the same first-frame image, pick a catalog id from the live model list, and send it with image_url. For frame-pinning detail see first and last frame on Seedance 2.5. Results will not match a Vidu render, since it is a different model.