Which Sume video models accept a reference video?

Nine Sume video ids accept a video input: seedance-2.5, seedance-2, seedance-2-fast, seedance-2-mini, wan-3.0, minimax-h3, minimax-h3-max and gemini-omni-flash-1.1, plus a Higgsfield motion-transfer model, which needs one. kling-3 and grok-imagine-video-1.5 do not take one.

A video input does three different jobs depending on the model, so read the mode before you choose. Facts are from Sume's catalog and video docs, read 2026-10-01.

What does the video do, and what are the limits?

Which model cannot run without a video?

The Higgsfield motion-transfer model needs one source video, plus 1 to 8 reference images, at 480p or 720p. The other video-input ids treat the video as optional.

Which one edits a video from a prompt?

Only gemini-omni-flash-1.1 has an edit mode: send video_url and a prompt to POST /v1/video-router/generate. It cannot be combined with frames or reference lists, and aspect_ratio is rejected.

curl -X POST https://api.sume.com/v1/video-router/generate \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: edit-001" \
  -d '{
    "model": "gemini-omni-flash-1.1",
    "prompt": "Make it night, keep the motion identical",
    "video_url": "https://example.com/clip.mp4",
    "resolution": "720p"
  }'