How to Make 15-Second Videos with Gemini Omni 1.2 Flash in Flow, Vids and the Gemini App

If you make short AI videos with Google’s tools, switch to Omni Flash 1.2 now: it generates and edits clips from 3 to 15 seconds long with sound, keeps faces and products more consistent, and lets you fix a clip with a follow-up message instead of starting over. Google released Gemini Omni 1.2 Flash on November 5, 2026 and made it generally available. The catch most people will hit first: in Google Flow, 15-second clips, Extend and upscaling need a Google AI Plus, Pro or Ultra plan, free accounts stay at 10 seconds, and Flow’s per-second credit pricing means a 10-second 720p clip now costs 30 credits instead of the 15 it cost on Omni Flash 1.1. This guide walks through Flow, Google Vids and the Gemini app step by step, shows what a clip really costs on each plan, and lists the gotchas, including the older models Google retires on December 5, 2026.

No coding is needed for anything in the main guide. There is a short optional appendix for people who use the Gemini API.

Last checked: November 6, 2026, against Google’s launch post, the Google Flow Help Center, the Google Vids help pages and the Gemini API pricing page. Google says Flow limits and credit costs “are subject to change at any time”, so always confirm the current cost in Flow’s prompt-box settings before you generate.

What changed in Omni 1.2 Flash, in plain words

Google’s launch post lists these changes compared with Omni 1.1:

  • Longer clips. You can generate and edit videos natively up to 15 seconds (up from 10), and pick the length in one-second steps from 3 to 15 seconds.
  • Edit in conversation. You can refine the same clip turn by turn, for example swapping the background or product, changing lighting or weather, or switching aspect ratio, without regenerating from scratch. Google says this works across full 15-second clips.
  • Better reference following. You can guide a clip with character, product or style images, short motion clips (up to three seconds) and audio cues, and the model holds faces, garments, packaging text and brand colors more consistently from first to last frame.
  • More natural speech with clearer lip-sync and distinct voices when several characters talk.
  • Steadier camera moves such as continuous drone-style flythroughs and zoom-outs with fewer unwanted jump cuts.
  • Cheaper on the API. Google cut the developer price by 30% versus Omni 1.1, to about $0.07 per second of 720p video.
  • Watermarked. Outputs carry an invisible SynthID watermark.

Google is also open about what still struggles: fast sports action, rapid script changes and complex physics. Plan around those rather than burning credits on them.

If you are new to Omni entirely, our earlier Gemini Omni Flash guide explains what an “any-to-video” model is and how the original release compared with other video tools.

Where you can use it today

According to Google’s launch post, Omni 1.2 Flash is rolling out across:

WhereWhoWhat you get
Google Flow (flow.google.com)Everyone with a Google account; more with a Google AI planText, frames and references to video, video editing, multi-turn edits; 15 s, Extend and upscaling for paid plans
Google VidsAnyone with a Google account or Google Workspace subscription15-second clips “at no additional cost” inside the Vids video editor
Gemini appRolling out to Google AI Ultra subscribers15-second videos
Google AdsAdvertisersVideo campaigns
Google AI Studio / Gemini API / Google CloudDevelopers and businessesPaid API, generally available

“Rolling out” means you may not see it on day one. If a feature below isn’t in your account yet, check again over the next few days rather than assuming you are doing something wrong.

Which plan do you need? (Google Flow)

Flow is where Omni 1.2 Flash is documented in the most detail, so start here. Google’s Flow models page splits features like this:

Without a Google AI subscription (free):

  • Text to Video, both aspect ratios, 3 to 10 seconds
  • Frames to Video (start/end frames), 3 to 10 seconds, currently web only
  • Ingredients/References to Video, 3 to 10 seconds
  • Video-to-video editing, 3 to 10 seconds

With Google AI Plus, Pro or Ultra:

  • All of the above at 3 to 15 seconds
  • Extend videos (Omni Extend): add 3 to 10 seconds, guided by up to 3 images, currently web only
  • Upscale Omni 360p videos up to 1080p, or 4K for Ultra subscribers

Everyone also gets two Omni-only extras: cheaper 360p drafts (one third of the 720p credit cost) and custom voices (pick a preset voice, then describe how to restyle it). Google notes some Omni features and voice options are only available in certain regions, and video editing itself is “only available in certain countries”.

How many credits a clip really costs

Flow charges credits per second for Omni 1.2 Flash. These are the official rates from Google’s Flow credits page:

ActionWho360p720p
Generate or edit, up to 10 sAll users1 credit/second3 credits/second
Seconds 11–15 of a clipPlus, Pro, Ultra2 credits per extra second6 credits per extra second
Extend by 3–10 sPlus, Pro, Ultra1 credit per second added3 credits per second added

Upscaling (Plus, Pro, Ultra): up to 1080p costs 0 credits for 1–10 seconds and 1 credit per extra second over 10. 4K (Ultra only) costs 50 credits flat for 1–10 seconds plus 5 credits per extra second.

Worked out from those rates, here is what common clips cost:

ClipCredits
10 s at 360p (draft)10
10 s at 720p30
15 s at 360p10 + (5 × 2) = 20
15 s at 720p30 + (5 × 6) = 60
Upscale a 15 s clip to 1080p5
Upscale a 15 s clip to 4K (Ultra)50 + (5 × 5) = 75

Two things that surprise people:

  1. Credits are per generation, not per request. Google warns that “a single request may generate 2 videos”. If your output count is set to 2, a 15-second 720p request costs 120 credits, not 60. Set outputs to 1 while you are experimenting.
  2. Editing is no longer a flat fee. On Omni Flash 1.1, any edit cost a flat 40 credits. On 1.2, Google says “edits cost the same credits per second as standard generations”, so a 5-second 360p fix is cheap and a 15-second 720p edit is 60 credits.

How far your credits go

Every account gets 50 Flow credits per day, whether or not you subscribe. Paid plans add a monthly pool on top:

PlanExtra Flow credits per month15 s 720p clips (60 credits each) from the monthly pool
Free0 (just the 50 daily)not available; 15 s needs a plan
Google AI Plus2003
Google AI Pro1,00016
Google AI Ultra ($100)10,000166
Google AI Ultra ($200)25,000416

On a free account, the 50 daily credits cover five 10-second 360p drafts, or one 10-second 720p clip with 20 credits left over. Daily credits refresh a day after your first generation and do not roll over; monthly plan credits refresh with your billing cycle and do not roll over either.

Free accounts have one more limit: during peak hours, roughly 2 PM to 5 PM UTC (10 PM to 1 AM Beijing time, 9 AM to noon US Eastern), Google says you may not be able to generate videos at all without a subscription. Images and the Flow Agent still work.

To check your balance, open Flow, click your profile picture at the top right, and look under your profile information.

Step by step: your first 15-second clip in Google Flow

These steps follow Google’s Flow editing help page. Flow works best in a desktop browser.

  1. Go to flow.google.com and sign in with your Google account.
  2. Open an existing project or start a new one.
  3. At the bottom of the prompt box, open your generation settings.
  4. Click Video. (Flow also makes images, so check you are not still on an image model.)
  5. Click the video model and choose Omni Flash 1.2.
  6. Set your preferences: orientation (landscape or portrait), number of outputs (choose 1 to save credits) and generation length (up to 15 seconds on a paid plan, 10 on free).
  7. Before you generate, open Settings at the top right of the prompt box to see the live credit cost. Google recommends checking the active model, resolution and current cost here every time.
  8. Type your prompt and click Generate.

A prompt template that suits 15 seconds

Fifteen seconds is long enough for a beginning, middle and end, so describe it as a short sequence rather than one frozen moment. Google’s own Vids guidance recommends including the subject, location, action, camera, lighting, dialogue, sounds and tone. For example:

A ceramic coffee mug on a sunlit kitchen counter. Shot 1: slow push-in as coffee is poured, steam rising. Shot 2: a hand picks up the mug and the camera follows it to a window. Shot 3: wide shot of a quiet street at sunrise. Warm, soft morning light, gentle acoustic music, sound of pouring coffee. No text or logos.

Tip: if you want readable words in the video (a product name, an offer), it is usually safer to add them later as a text layer in an editor such as Vids, where you control spelling. Google does say 1.2 holds packaging typography better, so try it, but check every frame.

Draft cheap, then finish

Google’s launch post describes the money-saving workflow directly: draft at 360p or 720p, fix details with a follow-up edit instead of regenerating, then upscale your best take. In Flow that means:

  1. Generate at 360p (20 credits for 15 seconds) until the motion and framing are right.
  2. Use a follow-up edit for small fixes instead of a full regeneration.
  3. Upscale the keeper to 1080p (5 credits for a 15-second clip) or 4K if you are on Ultra.

Edit an existing clip with follow-up prompts

You can edit both videos you generated in Flow and videos you upload.

  1. In the generation settings, choose Video, then Omni Flash 1.2 (steps 3 to 5 above).
  2. Upload, drag or select a video: up to 60 seconds and 1 GB, in .mov, .mp4, .avi or .wmv. Anything longer than 30 seconds must be trimmed to 30 seconds inside Flow first.
  3. Click the video. In the trimming window, select the segment to edit: up to 10 seconds, or up to 15 seconds on Plus, Pro or Ultra.
  4. Describe the change, for example “Change the lighting to a cinematic sunset” or “Add a text overlay that says ‘Coming Soon’” (both are Google’s examples).
  5. Optionally add ingredients (reference images), then click Generate.
  6. Keep refining with follow-up prompts. Flow’s help page says you can go up to 3 conversational turns without losing the context of earlier edits.

You never lose the original: the History panel on the right keeps every version and the prompt that made it. To reuse an older version in a scene, hover over it in History and click Save to Project.

You can also edit with the Flow Agent: turn on Agent in the prompt box, drag in one or more videos, describe the change and click Generate.

Make a clip longer with Extend

Extend is now an Omni 1.2 feature. Google retired the old Veo Extend on November 5, 2026, so to extend a clip in Flow you must be a Plus, Pro or Ultra subscriber using Omni Flash 1.2.

  1. In your project, click the clip you want to extend.
  2. At the bottom, click Extend.
  3. Select the extended duration (3 to 10 seconds).
  4. Describe how the action should continue. You can add up to 3 image references to keep characters and the environment consistent.
  5. Click Generate.

Extended clips cannot take other edit modes afterwards, such as insert, remove and camera, so do those edits before you extend. To build anything longer than one clip, use Scenebuilder: hover over a clip, click More > Add to Scene, drag clips into order, trim the ends with the handles, preview and download the scene.

Using it in Google Vids

Google Vids is Google’s browser video editor, and Google’s launch post says you can “generate 15-second clips at no additional cost in Google Vids with a Google account or Google Workspace subscription”. The Vids help page gives these steps:

  1. Open Google Vids on your computer (typing vids.new in the address bar works).
  2. From the start menu, select Create AI videos. In an existing video, select AI video on the right instead.
  3. In the side panel, select Create.
  4. Optional: select Ingredients and upload up to 7 reference images, or select Avatar to use an avatar (including its voice).
  5. Enter your prompt, including subject, location, action, camera, lighting, dialogue, sounds and tone.
  6. Optional: choose Landscape or Portrait below the prompt (it matches your canvas by default).
  7. Choose a duration.
  8. Click Generate, then Insert to add the clip to your video.

Gotchas specific to Vids:

  • The help page hasn’t caught up yet. As of November 6 it still lists durations of “between 3 and 10 seconds”. If you don’t see 15 seconds in the duration picker, the rollout hasn’t reached your account; check again later.
  • Insert before you close the tab. Google says the session history of generated clips disappears when you close the Vids tab, so insert anything you want to keep.
  • No sound? Google suggests checking your speakers, then making sure the generated clip’s track isn’t muted in the Vids timeline before you regenerate.
  • Monthly limits apply. The Vids help page says most users can generate up to 50 videos per month, resetting at 12 AM Pacific time on the first of the month, and that the limit is shared on Google AI Pro and Ultra family plans. Your exact allowance depends on your account and plan, so watch for Vids’ limit notices.

Using it in the Gemini app

Google says 15-second videos in the Gemini app are rolling out to Google AI Ultra subscribers. Google hasn’t published separate step-by-step instructions for the Omni 1.2 change in the app yet, so if you are on Ultra, look for it in the app’s video option over the coming days. If you are on Plus or Pro and want 15-second clips today, Flow is the documented route.

Whatever app you use, the same privacy settings apply to your prompts in Gemini; see our guide to stopping Gemini from training on your chats if that matters for client work.

December 5: older models are going away

Google’s Flow pages give two dates to plan around:

  • November 5, 2026: Veo Extend was retired. Extending now requires Omni Flash 1.2 on a paid plan.
  • December 5, 2026: Veo 3.1 Fast, Veo 3.1 Quality and Omni Flash 1.1 are retired in Flow. Until then they still work at their standard rates. Google recommends moving to Omni Flash 1.2 or Veo 3.1 Lite.

Veo 3.1 Lite stays as the low-cost option for simple 4, 6 or 8-second 720p clips: 10 credits per generation (5 for Ultra subscribers). It does not support video editing or Extend.

If you have a project mid-way on Omni Flash 1.1 or Veo 3.1 Fast, finish or re-create its key shots before December 5 so you are not caught out.

Troubleshooting and FAQ

I only see 10 seconds as the maximum length. In Flow, 15-second generation, editing and Extend require Google AI Plus, Pro or Ultra. On a paid plan, make sure Omni Flash 1.2 (not 1.1 or Veo) is selected. In Vids, the 15-second option is still rolling out.

Why did my clip cost twice what I expected? Check the number of outputs. Credits are charged per generated video, and one request can produce two.

Flow says I can’t generate video right now. On a free account this is likely the peak-hours limit (about 2–5 PM UTC). Try outside that window, or note that images and the Agent still work.

The Extend button is missing. Extend is Plus, Pro and Ultra only, on Omni Flash 1.2, and currently web only. Also remember that once a clip is extended, it can no longer take insert, remove or camera edits.

My character’s face changes between shots. Add a clear character image as an ingredient, keep the prompt describing the same person each turn, and use follow-up edits (up to 3 turns keep context) rather than fresh generations.

Is 10 seconds cheaper on 1.2 than on 1.1? Not in Flow. A 10-second 720p clip was 15 credits on Omni Flash 1.1 and is 30 on 1.2, while 360p went from 7 to 10. The 30% price cut Google announced applies to the Gemini API. In Flow, the gains are length, edits priced by the second instead of a flat 40, and quality.

Can people tell it’s AI? Every Omni 1.2 clip carries Google’s invisible SynthID watermark, so it can be identified as AI-generated. Label AI video where the platform you post to requires it.

Appendix (optional): Gemini API pricing

You don’t need this to use Flow or Vids. If you or your AI coding agent builds on the Gemini API, the model ID is gemini-omni-1.2-flash, and you can try it in Google AI Studio. Per Google’s Gemini API pricing page, it is paid tier only (no free tier), billed by output tokens at $12.00 per 1M video tokens, which works out to:

ResolutionApprox. price per second15 s clip
360p$0.023about $0.35
720pabout $0.07about $1.04
1080p (upsampled)$0.104about $1.56
4K (upsampled)$0.209about $3.14

Google’s figures are $0.07 per second and about $1.04 for a 15-second 720p clip; the other 15-second totals are the per-second price multiplied by 15. Input is $1.50 per 1M tokens for text, image or video. One date for API users: from December 3, 2026, all requests to gemini-omni-1.1-flash automatically route to gemini-omni-1.2-flash and are billed at 1.2 rates. If you’d rather not touch code, ask your AI agent to “switch our video generation calls from gemini-omni-1.1-flash to gemini-omni-1.2-flash and confirm the per-second cost”.

For the text side of Google’s lineup, see our guide on how to use Gemini 3.6 Flash.

Sources