How many jobs make a 3-minute vertical video: 6, 12 or 18 by model

The number of generation jobs for a 180-second vertical video is 180 divided by the model's maximum clip length: 6 on Wan 3.0 and Seedance 2.5 (30 s), 12 on MiniMax H3 (15 s) and 18 on Gemini Omni Flash 1.1 (10 s). Every job is billed separately and the Timeline join adds $0.30 for three render minutes.

180 seconds is the top of what a Short can be (YouTube Help, read 2026-10-09). If you only need 90 seconds, halve every count.

Clip count and cost at 180 s

All four list 9:16. The cost column is rate x 180 plus $0.30, from the Sume price tool on 2026-10-09.

Why the count matters

Each job has its own queue time, its own chance to fail, and its own prompt that must match the others. Timeline has a limit too: auto render chunks past 12 slots, so the 18-clip plan uses a chunked render and cannot ask for single.

# submit all jobs of one plan; poll each polling_url until completed
for i in 1 2 3 4 5 6; do
  curl -s -X POST https://api.sume.com/v1/videos \
    -H "Authorization: Bearer $SUME_API_KEY" -H "Content-Type: application/json" \
    -d "{\"model\":\"wan-3.0\",\"prompt\":\"Scene $i of 6: ...\",\"aspect_ratio\":\"9:16\",\"resolution\":\"720p\",\"duration\":30}"
done

Gotchas

• Seedance 2-class models stop at 15 seconds a job, which makes 12 jobs for the same 180 s; use seedance-2.5 for 30.
• Seedance 2.5 is token-priced; 0.5778 is the 720p 9:16 rate from the price tool.
• MiniMax H3 lists 480p and 768p; MiniMax H3 Max adds 1080p at the same 15 s limit.
• Jobs run asynchronously, so submit them in one burst and poll; a callback_url (HTTPS) can notify you on completion.