Summary: Gemini Omni 1.2 Flash bills video by output tokens at $12.00 per million, which Google converts to about $0.07 per second at 720p. A full 15-second 720p clip costs about $1.04 in output tokens, a 360p draft about $0.35, and 4K about $3.13. There is no free tier on the Gemini API, and from December 3, 2026 every request to the older Omni 1.1 Flash is routed to 1.2 and billed at the 1.2 rate.
Google announced the model on November 5, 2026 with a headline "30% lower price." That number is right, but it hides the parts that actually decide your bill: which resolution you render at, how many drafts you throw away, and how many edit turns you run. This breakdown turns the official token rates into per-clip and per-month numbers you can plan with.
The official prices, in plain terms
On the Gemini API pricing page, Gemini Omni 1.2 Flash (gemini-omni-1.2-flash) is listed for the paid tier only. The free tier column says "Not available" for both input and output.
- Input: $1.50 per 1M tokens (text, image or video).
- Output: $9.00 per 1M tokens for text, $12.00 per 1M tokens for video.
- Data use: on the paid tier, the "Used to improve our products" row says No.
Video output is counted at a fixed token rate per second, which depends on resolution. Google publishes these rates in the footnote under the table:
- 360p: 1,931 tokens per second (about $0.023/s)
- 720p: 5,792 tokens per second (about $0.07/s)
- 1080p with upsampling: 8,688 tokens per second (about $0.104/s)
- 4K with upsampling: 17,376 tokens per second (about $0.209/s)
The model page lists output clips of 3 to 15 seconds at 360p, 720p, 1080p or 4K, all at 24 FPS. The docs mark 1080p and 4K as upscaled output, and 720p is the default.
What one clip costs
Multiplying the token rate by $12 per million gives these output-only prices. We rounded to the cent; Google's own footnote quotes the same ~$0.70 for 10 seconds and ~$1.04 for 15 seconds at 720p.
- 15 s at 360p: 28,965 tokens, about $0.35
- 15 s at 720p: 86,880 tokens, about $1.04
- 15 s at 1080p: 130,320 tokens, about $1.56
- 15 s at 4K: 260,640 tokens, about $3.13
Two details from the docs change these numbers in practice:
- The default length is 10 seconds. If you do not set
durationinresponse_format, you get a 10-second clip, even if your prompt says "15 seconds." The duration field wins over the prompt text. Set it explicitly, as a whole number of seconds from 3 to 15 (for example"15s"). A value over 15 returns an invalid argument error. - Some modes cap at 10 seconds. First-and-last-frame interpolation and per-turn video extension support 3 to 10 seconds. Extension can grow a video to 30 seconds total, but each turn adds at most 10.
Editing turns are where the money goes
Omni 1.2's selling point is multi-turn editing: you keep the clip and ask for a new background, lighting or aspect ratio instead of re-prompting from scratch. The docs say each turn "produces a new video," and billing is based on total output tokens. Our working assumption is therefore that every edit turn is billed like a fresh clip of the same length and resolution. Google does not publish a separate edit price, so treat this as an estimate and check your usage after your first few turns.
Google's own advice in the launch post points the same way: draft at 360p or 720p, fix details with follow-up edits instead of regenerating, and only go to 1080p or 4K for the clips you keep. At these rates, a 360p draft costs about a third of a 720p take, so the cheapest workflow is to explore at 360p and render only the winner at higher resolution.
Monthly budget for a Shorts channel
Assume 30 vertical videos a month (set aspect_ratio to "9:16"; 16:9 is the default), each 15 seconds:
- One 720p take per Short: 30 x $1.04 = about $31 a month.
- One 360p draft, one 720p take and one 720p edit per Short: $0.35 + $1.04 + $1.04 = about $2.43 each, about $73 a month.
- Same single take at 1080p: 30 x $1.56 = about $47 a month.
Input tokens ($1.50 per million) come on top. For a short text prompt they are minor, but reference images and uploaded video also count as input, and Google does not list a per-image or per-second input conversion on the pricing page, so we have not put a number on them.
Omni 1.2 vs 1.1 vs Google Vids
| Option | Video price | Max clip length | Free use | Notes |
|---|---|---|---|---|
| Gemini Omni 1.2 Flash (API) | $12.00 per 1M output tokens, about $0.07/s at 720p | 15 s (extend to 30 s total) | No free tier on the API | Generally available; 360p to 4K; multi-turn editing up to 15 s clips |
| Gemini Omni 1.1 Flash (API) | $17.50 per 1M output tokens, about $0.10/s at 720p | 10 s | No free tier on the API | From Dec 3, 2026 requests route to 1.2 and are billed at 1.2 rates |
| Google Vids (app) | "No additional cost" per Google's launch post | 15 s clips | Included with a Google account or Google Workspace subscription, per Google | App, not API; for people who do not need automation |
The switch from $17.50 to $12.00 per million is a 31% cut at the same tokens per second, which matches Google's "30%" claim. If you still call gemini-omni-1.1-flash, you do not need to change code to get the lower price after December 3, but you will also be getting a different model, so re-test prompts that worked on 1.1.
Editor's note
Who it's for: developers and small teams generating short ads, product spins or explainers in volume, who want predictable per-second pricing and an API.
Who it's not for: anyone who wants to try AI video for free through the API (there is no free tier), or people making a few clips a month, who can use Google Vids instead.
Gotchas we verified in the docs:
- Forgetting
durationgives you a 10-second clip. Writing "15 seconds" in the prompt does not change that. - Setting
store=falsespeeds things up, but the docs say the video then cannot be edited in later turns withprevious_interaction_id. Keep storing clips you might edit. - For videos over 4 MB, the docs recommend
delivery="uri"to avoid payload size limits. - Editing or extending uploaded videos is not available in the EEA, Switzerland and the UK; editing videos the model generated is supported there.
- Uploading audio references is not supported in the current API version.
Related reading on lightmind24: for the language-model side of Google's lineup, see Gemini 4 Carbon Pricing: Intro Rates End Nov 30, vs Argon. For another per-task cost breakdown, see Claude Fable 5.2 Pricing, and if you work in Google's developer tools, our Antigravity CLI setup guide.
Screenshots of Google's launch post, the Gemini API pricing and model pages, and our own cost table are at the end of this post.
FAQ
Is Gemini Omni 1.2 Flash free to try?
Not on the Gemini API: the pricing page lists the free tier as not available. Google's launch post says you can generate 15-second clips at no additional cost in Google Vids with a Google account or Workspace subscription, and that it is rolling out in the Gemini app to Google AI Ultra subscribers.
How much does a 15-second video cost?
About $1.04 in output tokens at 720p, $0.35 at 360p, $1.56 at 1080p and $3.13 at 4K, based on Google's published tokens per second and $12 per million video output tokens. Input tokens are extra.
What happens to Omni 1.1 Flash?
From December 3, 2026, requests to gemini-omni-1.1-flash automatically route to gemini-omni-1.2-flash and are billed at the 1.2 rates, according to the pricing page and model docs.
Draft at 360p, set the duration explicitly, and pay for 720p or above only for clips you keep.
Last verified: 2026-11-17
Sources:
Google Blog: Gemini Omni 1.2 Flash
Gemini API pricing
Gemini API: Gemini Omni Flash model page
Gemini API docs: Generate and edit videos with Gemini Omni Flash
Image credits: screenshots of the official Google blog and Gemini API documentation; cost table by lightmind24.
lightmind24 is an independent review site, not affiliated with Google.
댓글
댓글 쓰기