Split a 16 or 20 second Sora shot into Omni clips and join them
A Sora shot of 16 or 20 seconds becomes two or three Gemini Omni Flash clips of 10 seconds or less on Sume, joined with Timeline 1.0. How to cut the beats.

To replace a 16 or 20 second Sora shot with Gemini Omni Flash 1.1, write it as two or three beats of 10 seconds or less, generate each beat as its own clip, and join the clips with a Timeline 1.0 render. The longest single Omni output Google documents is 10 seconds, so there is no one-request equivalent of a 20 second Sora generation.
OpenAI's Sora 2 Prompting Guide lists clip lengths of 4, 8, 12, 16 and 20 seconds, and its video API guide says both Sora 2 models support 16 and 20 second generations. Google's Omni documentation lists 3 to 10 second output. Both were read on 2026-10-03, and OpenAI's guide now states that the Sora 2 models and Videos API were shut down on September 24, 2026.
Cut the beats before you write the prompts
Sora's guide describes motion as clear beats, for example four steps to a window, a pause, then a curtain pulled in the final second. Use those beats as your cut points. A 20 second shot with four beats of five seconds each becomes two Omni clips of 10 seconds, each holding two beats; a 16 second shot with two beats of eight seconds becomes two clips of 8 seconds.
Keep the camera, subject and light wording identical in every clip's prompt so the joins look like one scene. For continuity of the starting pose, Google's Omni documentation supports first and last frame control; on Sume that is image_url plus an optional end_image_url in the same request envelope.
Cost of the split
Google lists Omni output at about $0.10 per second at 720p, billed on tokens at 5,792 tokens per second of 720p video. Sume's docs say billing is the provider list rate times 1.25 per output second by resolution, so at 720p the same figure becomes $0.125 per second. The table works that out for the common splits.
| Sora shot | Omni clips | Seconds billed | Google list at $0.10/s | Sume at list x 1.25 |
|---|---|---|---|---|
| 16 s | 2 x 8 s | 16 | $1.60 | $2.00 |
| 20 s | 2 x 10 s | 20 | $2.00 | $2.50 |
| 20 s | 3 x 7 s (one spare second per join) | 21 | $2.10 | $2.63 |
Join the clips on Timeline 1.0
Timeline 1.0 takes one audio spine plus ordered video[] slots and returns one MP4. Each slot has source_url, an on-spine start, a duration, and an optional transition on slots after the first (fade, dissolve and so on, 1 second at most). The docs price a render at $0.10 per output minute rounded up, so a 20 second join costs one minute.
Omni clips carry their own synced audio. If you want that sound under a continuous bed, detach it or use audio.mode: "silence" and a soundtrack; read the Timeline docs for the exact fields before you pick one, because the audio spine decides the output length. Every URL must already be a media.sume.com artifact, which generated clips are.
A join request
The request below joins two 10 second clips with a quarter-second dissolve. Replace the URLs with your own clip artifacts. audio.duration_seconds sets the output length.
curl -X POST https://api.sume.com/v1/timeline-1.0/render \
-H "Authorization: Bearer $SUME_API_KEY" \
-H "Content-Type: application/json" \
-H "Idempotency-Key: sora-split-join-001" \
-d '{
"audio": { "mode": "silence", "duration_seconds": 20 },
"video": [
{ "source_url": "https://media.sume.com/artifacts/artf_demo/beat1.mp4", "start": 0, "duration": 10 },
{ "source_url": "https://media.sume.com/artifacts/artf_demo/beat2.mp4", "start": 10, "duration": 10,
"transition": { "type": "dissolve", "duration": 0.25 } }
]
}'Sources
Related posts
More in Models
- Stable Audio Open 1.0: 47-second limit and the Community License
Stable Audio Open 1.0 makes up to 47 seconds of 44.1 kHz stereo audio, trained on Freesound and FMA. Read the license note, limits, and a Sume alternative.
- Sume auto or a pinned video model: what changes past 10 seconds
Auto video defaults to 720p and 8 seconds with 3 to 10 second clips. For 12, 15 or 30 seconds, 1080p or 21:9, pin a model id. How to decide on Sume.
- Is Sume's kling-3 Kling VIDEO 3.0 or Omni 3.0? Read the catalog row
Sume's catalog names kling-3 'Kling Video v3 Pro' and lists no reference inputs. Kling describes Omni 3.0 as reference-driven, so match by capability.
- Sume Voices page lists 16 languages, Sonic 44: which list applies?
The Sume Voices page offers 16 languages when you create a voice; Cartesia states 44 for Sonic. The TTS request language field is a BCP-47 code you set per job.
Written by Sume