Gemini Omni Flash 1.1 at $0.125 a second on Sume vs $0.10 at Google
Sume bills Omni Flash 1.1 at $0.125 per second at 720p. Google lists about $0.10. Here is the 25% gap in dollars for 100 holiday clips, and what it covers.
Gemini Omni Flash 1.1 costs $0.125 per second at 720p on Sume and about $0.10 per second on Google's own Gemini API. On 100 ten-second holiday clips that is $125.00 against roughly $100.00, a $25.00 difference.
This post is about that gap, since it is the first thing a buyer with a Q4 budget will notice. It is a margin, and it is stated in Sume's docs, not hidden.
The two price lines
Google's Gemini API pricing page prices Omni Flash video output at $17.50 per million video tokens, which it states is approximately $0.10 per second of 720p video. Sume's Video Router catalog row for the same model records a provider list of $0.10 per second at 720p and then bills list × 1.25.
Arithmetic: $0.10 × 1.25 = $0.125 per second. For a 10-second clip, 10 × $0.125 = $1.25, against 10 × $0.10 = $1.00 direct.
| Route | Per second | 10 s clip | 100 clips |
|---|---|---|---|
| Google Gemini API, approx. | $0.10 | $1.00 | $100.00 |
| Sume Video Router (list x 1.25) | $0.125 | $1.25 | $125.00 |
| Difference | $0.025 | $0.25 | $25.00 |
What the extra $0.25 a clip buys
Only claims that the repo supports belong here. A Sume job has an Idempotency-Key, so a retried request returns the original job and does not bill twice. It has a signed webhook through callback_url, a status route at /v1/jobs/{id}/status, and one workspace USD balance that also pays for avatar, music, TTS and Timeline calls.
If your pipeline only ever calls this one model and you already hold a Google account with billing, the direct route is cheaper per second. If the clip is one step in a chain (voice, music, captions, assembly), the shared balance and job lifecycle can matter more than $0.25.
Same model, other resolutions
The gap scales with the rate. At 1080p the catalog list is $0.15, so Sume bills $0.1875 per second. At 4K the list is $0.30 and Sume bills $0.375. Only compare the Google figure for a resolution you read on Google's page yourself, because this post reads only the 720p approximation there.
A decision rule for a Q4 batch
Whichever route you pick, price one batch with real clip lengths first. Eight seconds instead of ten changes the total by 20%, which is larger than the margin discussed here.
- Under 200 clips, single model, existing Google billing: go direct and keep the $25 per 100 clips.
- A chain with voice, music or captions, or several teammates sharing one balance: use Sume and accept the margin.
- Edits of a finished clip (swap the gift, keep the shot): the Sume catalog exposes
video_urledit mode on this model, so the same model id covers both generation and edit.
Sources
Related posts
More in Comparisons
- Gemini TTS: 30 prebuilt voices or 150+? Which number to quote
Google's docs say 30 prebuilt voices; its changelog says 150+ prebuilt and custom. Quote the one that fits and compare with Sume voice-id selection.
- Two-voice dialogue: Gemini TTS vs ElevenLabs Text to Dialogue
Gemini TTS handles up to 2 speakers per request; ElevenLabs Text to Dialogue keeps continuity with surrounding text and request IDs. In Sume you stitch jobs.
- Voice Design from a text description: Gemini vs ElevenLabs
Gemini TTS has Voice Design and 200 custom voices per project; ElevenLabs lists Voice Design on v4. Sume's docs cover cloning, not voice design.
- GitHub Actions schedule or a Sume schedule for a weekly video run?
For a weekly video run, GitHub Actions cron is UTC-first and best-effort at busy times; a Sume schedule has a timezone and a spend cap. You can also chain them.
Written by Sume