A 3-minute YouTube Short from Omni clips: 18 takes, what it costs

A full three-minute Short built from Gemini Omni Flash clips is 18 clips of 10 seconds. Through Sume that is $22.50 at 720p or $33.75 at 1080p, plus $0.30 to join them in Timeline 1.0 at $0.10 per output minute.

YouTube's help page says Shorts can be up to three minutes and square or vertical (read 2026-10-08). Omni's ceiling is 10 seconds per request, so the length is a count of clips.

Cost by clip length

Shorter shots mean more clips for the same 180 seconds but the same video cost, since billing is per output second.

Practical plan

• Write the script as 18 shots, each prompt standing alone. Omni generates its own audio per clip, but a Timeline render is built around one audio spine, so plan the sound track first: send one spine as audio.url or audio.parts[], or use silence mode.
• Draft all 18 at 360p first: $6.75 for the set.
• Re-run only the approved shots at 720p or 1080p, then join.

Limits

Each clip is its own job with its own reservation, so an 18-clip run needs the balance for the jobs in flight. Concurrency is capped per workspace; if 429 queue_full appears, submit in waves.

The join

A Timeline 1.0 document for 180 seconds sets audio.duration_seconds to 180 and lists 18 video[] slots whose start values step by 10 from 0 (video[0].start must be 0). The reserve is ceil(180 / 60) = 3 minutes, so $0.30. POST /v1/timeline-1.0/plan compiles the document without a job or a reserve, so check it before rendering. Timeline accepts spines up to 1800 seconds, so 180 seconds is well inside the limit.