A 60-second Short as twenty 3-second Omni cuts: $7.50 at 720p

Gemini Omni Flash 1.1 accepts as little as 3 seconds a job, so a fast-cut 60-second vertical Short can be twenty separate 3-second clips. They cost $7.50 at 720p and $11.25 at 1080p (the rate is linear, so six 10-second clips cost the same), plus $0.10 for the one-minute Timeline render.

Fast cuts suit Shorts, which are vertical and up to 3 minutes (YouTube Help, read 2026-10-09). The catch is operational: twenty jobs, twenty slots and a render that Timeline will chunk.

Cost

Omni Flash lists 360p, 720p, 1080p and 4K in 16:9 or 9:16, 3 to 10 seconds a job.

Building the body

Timeline takes 1 to 200 slots. With render.strategy auto it chunks past 12 segments, and it refuses single above 12 (render_strategy_unsafe). More than 8 adjacent fades fail with too_many_chained_transitions, so the sample puts a fade on every third join and hard-cuts the rest.

import json
slots = []
for i in range(20):
    slot = {"source_url": f"https://media.sume.com/artifacts/artf_demo/cut{i + 1:02d}.mp4",
            "start": i * 3, "duration": 3}
    if i and i % 3 == 0:          # a fade on every third join, hard cuts between
        slot["transition"] = {"type": "fade", "duration": 0.25}
    slots.append(slot)

body = {"audio": {"mode": "silence", "duration_seconds": 60},
        "video": slots, "render": {"strategy": "auto"},
        "output": {"width": 1080, "height": 1920}}
print(json.dumps(body))

Limits

• Twenty prompts means twenty places for the style to drift; pass the same reference image to every request.
• A cut shorter than 3 seconds is not possible from Omni; use Timeline's duration (min 0.2 s) to shorten a clip inside the render instead.
• Plan first with POST /v1/timeline-1.0/plan, which is unbilled, to see segment_count and billable_minutes.
• Each slot start must increase and the first must be 0.