Listing photo to a 10-second tour clip: MiniMax H3 at 768p is $0.75

One listing photo turns into a 10-second tour clip for $0.75 on MiniMax H3 at 768p ($0.075 per second), or $1.00 on H3 Max at the same resolution. A listing with 12 photos costs 12 x $0.75 = $9.00 on H3.

Both models take a start image, run 5-15 seconds, and always produce native audio. Prices below come from the Sume price tool (read 2026-10-09).

Price of a 10-second tour clip

The cheapest row that still looks like a listing video is H3 at 768p. H3 Max costs 33% more per second at the same resolution, so use it only on the photos where the first H3 take missed the prompt.

The request body

Send the photo as image_url on the Video Router and ask for one camera move. 768p is native on MiniMax, so send 768p, not 720p.

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: listing-4417-tour-v1" \
  -d '{
    "model": "minimax-h3",
    "prompt": "Slow push-in through the living room, soft daylight, no people",
    "image_url": "https://example.com/living-room.jpg",
    "resolution": "768p",
    "duration": 10,
    "aspect_ratio": "16:9",
    "mode": "async"
  }'

Shorter clips cost less

MiniMax bills each second, so a 5-second clip per photo costs half of a 10-second one. Pick the length by how long the photo needs to hold the viewer, not by the maximum.

Limits and gotchas

Read capabilities from GET /v1/video-router/models if you build a form for this, so the duration and resolution menus come from the catalog.

• Duration must be 5-15 seconds. A 4-second request is refused.
• H3 has no 1080p row. For 1080p, use H3 Max.
• There is no generate_audio toggle, so each clip has a stereo soundtrack you may need to mute in your editor.
• Keep people out of the prompt unless the photo has them: the model animates what the photo shows.