Gemini Omni video is Google’s AI video model inside the Gemini app, Flow, Google Vids and the Gemini API, and it turns a text prompt, up to five photos or a short clip into a finished video with sound. On 5 November 2026 Google made Gemini Omni 1.2 Flash generally available, raising the maximum clip length from 10 to 15 seconds, improving character consistency and cutting the API price by 30%. This guide shows you exactly how to create, prompt, edit and extend Gemini Omni video, what each plan includes, and the limits worth knowing before you rely on it.
Everything below is checked against Google’s own announcement, Help Center and developer documentation as of 7 November 2026. Where Google has not published a number (daily limits, for example), we say so rather than guess.
Hands-on test results: coming soon
We have not generated Gemini Omni clips ourselves yet. Every fact in this guide is checked against Google’s official documentation, and we will add our own test results and screenshots as soon as our test account has video access.
In this article
What is Gemini Omni video?
Gemini Omni is a model family that Google DeepMind introduced on 19 May 2026. Its first output type is video: it can combine text, images, video and audio cues as inputs and return a short clip with synchronised audio. The headline difference from earlier tools is conversation. Instead of rewriting a long prompt every time, you edit a Gemini Omni video by chatting (“make it night”, “swap the mug for a laptop”) and the model keeps everything else consistent.
You can use Gemini Omni video in several places:
- Gemini app (web, Android and iOS) through the Create video tool.
- Google Flow, Google Labs’ filmmaking tool, for directing scenes and multi-turn edits.
- Google Vids, which Google says offers 15-second clips at no additional cost with a Google account or Workspace subscription.
- The Gemini API and Google Cloud for developers and businesses, now generally available.
What’s new in Gemini Omni 1.2 Flash?
Gemini Omni 1.2 Flash is the version Google released on 5 November 2026. Here is how it compares with the previous 1.1 Flash model, using Google’s published figures.
| Specification | Gemini Omni 1.1 Flash | Gemini Omni 1.2 Flash |
|---|---|---|
| Maximum clip length | 10 seconds | 15 seconds (3 to 15, in 1-second steps) |
| API price, 720p | 30% higher than 1.2 | $0.07 per second |
| Draft resolutions | 360p or 720p | 360p or 720p |
| Upscaled output | 1080p or 4K | 1080p or 4K |
| Status | Previous version | Generally available (5 November 2026) |
Google lists these improvements in Gemini Omni 1.2 Flash:
- Better reference adherence: combine character, product and style images, short motion clips (up to 3 seconds) and audio cues, with faces, garments, typography and brand colours staying consistent.
- Less drift: stronger consistency of characters, objects and scenes from the first frame to the last, plus first and last keyframe interpolation and clip extension.
- Multi-turn editing: swap products or backgrounds, change lighting or weather, or switch aspect ratio without regenerating.
- Speech: more natural dialogue, clearer lip-sync and multiple speakers.
- Camera control: complex moves such as FPV flythroughs, zooms and orbits with fewer unwanted jump cuts.
Every Gemini Omni video also carries an invisible SynthID watermark so it can be identified as AI-generated.
Who can use Gemini Omni video, and on which plan?
In the Gemini app, Gemini Omni video requires a paid Google AI plan on a personal account, or a qualifying Workspace licence on a work or school account. It is not available to users under 18. Google’s limits page lists video generation for these tiers:
| Plan | Video generation | Usage limit (relative) | Max clip length in the app |
|---|---|---|---|
| Free (no AI plan) | Not included | Not applicable | Not applicable |
| Google AI Plus | Included | 2x standard | 3 to 10 seconds |
| Google AI Pro | Included | 4x standard | 3 to 10 seconds |
| Google AI Ultra | Included | 5x or 20x AI Pro, depending on subscription | Up to 15 seconds (1.2 rolling out) |
Google does not publish a fixed number of videos per day. Limits are measured in compute, refresh every five hours and weekly, and video is the most expensive request type. You can see your remaining usage at gemini.google.com under Settings > Usage Limits. Prices for each plan vary by country, so check the current price in your region before subscribing.
Region note: uploading your own videos for editing, and creating or sharing avatars, is not available in the EEA, Switzerland, the UK and some US states. You can still generate a Gemini Omni video from text and photos there, and you can edit videos the model generated for you.
Which Google tool should you use for Gemini Omni video?
The same model sits behind several Google products, and each suits a different job. Pick the one that matches what you are making.
| Where | Best for | What you need |
|---|---|---|
| Gemini app | Quick clips, chat edits, avatars, sharing to YouTube | Google AI Plus, Pro or Ultra (Ultra for 15-second clips) |
| Google Flow | Directing scenes and multi-turn edits for longer creative projects | Access to Flow through Google Labs |
| Google Vids | Short clips inside work videos, training and presentations | A Google account or Workspace subscription (no additional cost, per Google) |
| Gemini API | Generating Gemini Omni video inside your own app or pipeline | A Gemini API key, billed per second of video |
| Google Cloud | Business use at scale with enterprise security | A Google Cloud project; Provisioned Throughput available |
If you are just starting, use the Gemini app: it has templates, the simplest editing loop and direct download. Move to the API only when you need to generate Gemini Omni video automatically or in bulk.
How do you create a Gemini Omni video in the Gemini app?
These steps follow Google’s Help Center instructions for the desktop app. The Android and iOS apps use the same Create video tool.
Step 1: Open the Create video tool
Go to gemini.google.com and sign in. Click Open sidebar at the top left, then click Create video. You can optionally pick a template to start from.
Step 2: Write your prompt
Type what should happen in the clip. Describe the shot, the subject, the action, the setting and the mood (the prompt section below gives a formula).
Step 3: Add references (optional)
Click Add image to attach up to five photos, or Add video to attach one clip. To appear in the video yourself, type @ followed by your Google username to use your avatar, or tag someone who has shared their avatar with you.
Step 4: Choose the aspect ratio
Text-only prompts default to landscape. If you upload a reference image or video, the Gemini Omni video matches its aspect ratio. Choose vertical before generating if the clip is for Shorts, Reels or TikTok.
Step 5: Generate and wait
Click Submit. Generation can take a few minutes. You cannot use the same chat while it renders, but you can open a new chat and come back.
Step 6: Download or share
Under the finished clip, click Share, then Download video or Share on YouTube. Review the clip before publishing it anywhere.
How do you write prompts that work for Gemini Omni video?
Google DeepMind’s official prompt guide recommends mixing five elements in every prompt:
- Shot framing and motion: close-up, wide shot, push in, orbit, FPV flythrough.
- Style: cinematic, documentary, stop-motion, webcam style, film camera.
- Lighting: golden hour, neon, soft window light, overcast.
- Location: where the scene takes place.
- Action: what happens, in order.
Useful camera phrases from the guide include “one continuous shot” (or “oner”) for no cuts, “static” or “locked off” for a fixed camera, and “push in”, “punch in” or “dolly zoom” for movement. For real-world scenes, you can be less prescriptive than with older video models: describe the outcome and let Gemini Omni fill in realistic detail.
Example prompt:
One continuous shot, slow push in. A barista in a small corner cafe pours latte art
into a white cup at golden hour. Warm window light, shallow depth of field,
documentary style. Soft cafe ambience and the hiss of the steam wand.Example prompt for a product clip (with a reference photo attached):
Use the attached photo as the product. Static, locked off shot on a clean wooden desk.
Morning light from the left. The bottle slowly rotates 180 degrees while steam rises
from a cup beside it. Minimal, premium commercial style. No text on screen.Notice that the product prompt states what to avoid (“No text on screen”) inside the prompt itself. Gemini Omni has no separate negative prompt field, so this is the only way to rule things out.
For dialogue, write the spoken line in quotes inside the prompt. For on-screen text, say what the text is, where it sits and how it animates. If you have a storyboard, upload it as an image and tell Gemini Omni to “follow the story exactly in order starting top left”.
How do you edit and extend a Gemini Omni video?
Editing is where Gemini Omni video stands out. After the first clip renders, reply in the same chat with one change at a time:
- “Change the camera angle to over the barista’s shoulder.”
- “Make it a rainy evening, keep everything else the same.”
- “Replace the white cup with a black ceramic mug.”
- “Switch to vertical 9:16.”
Each turn produces a new version and keeps what already works. You can attach new image or video references on any turn. Adding “keep everything else the same” is the simplest way to stop unwanted changes.
To make a clip longer, ask Gemini what happens next. In the Gemini app you can extend a video by up to another 10 seconds at a time, and Google’s current API documentation caps an extended video at 30 seconds in total (its older prompt guide, written for Omni 1.1, mentions 40). Extensions add to the end of the clip; they do not insert footage in the middle.
How do developers use Gemini Omni video through the API?
Gemini Omni 1.2 Flash is generally available on the Gemini API and Google Cloud, with Provisioned Throughput on Google Cloud. Google’s developer documentation lists the model as gemini-omni-1.2-flash, and Google’s announcement puts the price at $0.07 per second of 720p video. At that rate a 10-second 720p Gemini Omni video costs about $0.70 and a 15-second clip about $1.05, before any upscaling.
The API uses the Interactions endpoint in the Google Gen AI SDK. This minimal Python example follows the pattern in Google’s documentation:
import base64
from google import genai
client = genai.Client() # reads your GEMINI_API_KEY environment variable
interaction = client.interactions.create(
model="gemini-omni-1.2-flash",
input="A marble rolling fast on a chain reaction style track, continuous smooth shot."
)
with open("marble.mp4", "wb") as f:
f.write(base64.b64decode(interaction.output_video.data))Google’s developer documentation now uses gemini-omni-1.2-flash in its examples. The older gemini-omni-1.1-flash ID still works during a transition window, and from 3 December 2026 requests to it are routed to 1.2 automatically. Other details from the documentation:
- Aspect ratio: 16:9 (default) or 9:16.
- Resolution: 360p, 720p (default), or upscaled 1080p and 4K.
- Editing: pass
previous_interaction_idto edit the last result; storage must stay enabled for this to work. - Inputs: you can upload videos of up to 15 seconds to edit, and add up to 6 reference images plus up to three reference videos of 3 seconds each (one reference video for clips longer than 10 seconds).
- Not supported: system instructions, temperature and a separate negative prompt field, so put anything you want to avoid into the prompt itself.
What are the limits of Gemini Omni video?
Gemini Omni video is impressive, but Google is open about where it still struggles. Plan around these points:
- Hard physics and fast action: Google says it is still improving fast sports action, rapid scene changes and complex physics.
- Short clips: 10 seconds on most plans, 15 seconds on AI Ultra and the API, up to 30 seconds with extensions on the API.
- No fixed daily quota: usage is compute-based and can be reduced during high demand.
- Region restrictions: uploaded-video editing and avatars are unavailable in the UK, EEA and Switzerland.
- English-first: the API documentation says non-English prompts have not been evaluated.
- Watermarking: every output carries SynthID, and Gemini can check whether a file was made with Google AI.
Common mistakes when making Gemini Omni video
1. Writing vague prompts
“A cool product video” gives the model nothing to work with. Name the shot, the style, the light, the place and the action, and your first Gemini Omni video will be much closer to what you pictured.
2. Regenerating instead of editing
If one detail is wrong, fix it with a follow-up instruction. Starting over throws away everything that worked and uses more of your limit.
3. Changing several things at once
Ask for one change per turn and add “keep everything else the same”. Multiple edits in one message make it harder for the model to preserve consistency.
4. Skipping reference images
For products, logos or a recurring character, upload clear reference photos. References are the most reliable way to keep a face, garment or brand colour identical across shots.
5. Publishing without a review
Google’s own guidance is to review generated videos before you rely on or publish them. Check hands, text, product details and lip-sync, and never use someone’s likeness without permission.
Real-world use cases for Gemini Omni video
- Social clips: vertical 9:16 Gemini Omni video for Shorts, Reels and TikTok, generated in minutes.
- Product demos: upload product photos as references and show the item in different settings without a shoot.
- Ad variations: generate one scene, then use chat edits to test different backgrounds, lighting or calls to action.
- Training and explainers: Google Vids users can add short Gemini Omni clips to internal videos.
- Storyboards and pitches: turn a sketched storyboard into a moving preview before committing to production.
- Prototyping for developers: use the API to generate video assets inside your own app or content pipeline.
Frequently asked questions
Is Gemini Omni video free?
Not in the Gemini app: you need a Google AI Plus, Pro or Ultra plan or a qualifying Workspace licence. Google says Google Vids offers 15-second Gemini Omni clips at no additional cost with a Google account or Workspace subscription, and at launch Google made Omni available for free in YouTube Shorts and the YouTube Create app.
How long can a Gemini Omni video be?
In the Gemini app, 3 to 10 seconds, or up to 15 seconds with AI Ultra. Through the API, Gemini Omni 1.2 Flash generates up to 15 seconds. Through the API, extensions can take a clip to 30 seconds in total.
How much does Gemini Omni video cost on the API?
Google lists Gemini Omni 1.2 Flash at $0.07 per second for 720p video, 30% less than 1.1 Flash. A 10-second 720p clip is therefore about $0.70. Check Google’s pricing page for other resolutions.
Does Gemini Omni video have a watermark?
Yes. Every output carries an invisible SynthID watermark. Google’s pages do not mention a visible watermark.
Can I use Gemini Omni video in the UK?
Yes, with a paid plan, you can generate videos from text and photos and edit the videos Gemini creates. Uploading your own videos for editing and creating avatars are currently unavailable in the UK, the EEA and Switzerland.
What is the difference between Gemini Omni and Veo?
Both are Google video models. Gemini Omni is built around conversational editing and mixing many input types in one request, and Google’s prompt guide notes you can be less prescriptive with Omni than with Veo for real-world scenes.
Getting started with Gemini Omni video
The fastest way to learn Gemini Omni video is to make one short clip with a five-element prompt, then improve it with three or four single-change edits. That loop teaches you more about the model than any long prompt.
If you are building a Google-based workflow, pair it with our guide to Gemini Notebook for source-grounded research to script your video, and see our Claude vs ChatGPT vs Gemini comparison to choose the right assistant for the writing side. You can find more guides on our tutorials page.
Sources: Google: Gemini Omni 1.2 Flash announcement, Gemini Apps Help: Generate videos, Gemini Apps Help: limits and plans, Gemini API: Gemini Omni Flash docs, Google DeepMind: Gemini Omni prompt guide.