Overview
Gemini Omni is Google DeepMind's new family of multi-modal generative video models, launched on May 19, 2026 with the first member — Gemini Omni Flash — rolling out the same day. Unlike Gemini 3.x Flash and Pro, which are language and agentic models, Omni is built for video creation and editing: turning image, video, text, and supported audio references into edited clips through natural, conversational instructions.
Among the best AI video generators, Gemini Omni is positioned for creators, storytellers, and editors rather than developers and analysts. It is rolling out through the Gemini app and Google Flow for Google AI Plus, Pro, and Ultra subscribers, and at no cost in YouTube Shorts and the YouTube Create app starting this week, with an API for developers and enterprises planned to arrive within weeks. All generated video carries Google's invisible SynthID watermark for provenance. It joins Google's other generative video model, Veo 3.1, in the family.
Key Features
- Multi-modal input compositing — Mix any image, video, audio (voice reference), and text prompt into a single video output, letting creators blend references in one request instead of stitching across separate tools.
- Conversational step-by-step editing — Refine scenes through natural-language conversation: change details, swap environments, adjust camera angles, or restyle without re-prompting from scratch.
- Physics-aware generation — Google says Omni has an improved intuitive understanding of gravity, kinetic energy, and fluid dynamics, helping produce more realistic motion and physical interactions and reducing the "AI-warped" feel common in earlier video models.
- Knowledge-grounded scenes — Combines physics intuition with Gemini's knowledge of history, science, and cultural context, so short prompts can produce more context-aware historical, scientific, or explanatory scenes, though outputs still need human review.
- Motion transfer from any reference — Apply pose and motion from one input video onto a character in a reference image, useful for character animation, dance choreography, and sports re-targeting; see our AI animation video generators roundup for adjacent tools.
- SynthID watermarking on every clip — Every generated video carries Google's invisible SynthID watermark, enabling downstream platforms and creators to verify AI provenance.
Pricing & Plans
Gemini Omni is gated behind Google's consumer AI subscriptions and YouTube surfaces; per-token API pricing has not been disclosed at launch.
| Access path | Cost |
|---|---|
| YouTube Shorts + YouTube Create app | At no cost, rolling out starting this week |
| Gemini app + Google Flow | Included with Google AI Plus, Pro, or Ultra subscriptions; access limits vary by tier |
| Developer / Enterprise API | Coming "in the next few weeks"; pricing tier not yet published |
Features and quotas vary by subscription tier and geography. Google has not released free-trial specifics for the Gemini-app paths beyond the existing trial mechanics of Google AI subscriptions; confirm current details on the official Google AI plans page.
Best For
- Creators who want to mix image, audio, and video references in one editing session without context-switching across tools.
- Storytellers experimenting with physics-accurate or knowledge-grounded short videos (period scenes, science explainers).
- YouTube Shorts creators looking for a free, in-app AI video generator with built-in watermarking.
- Animation and motion designers needing pose / motion transfer from a reference clip onto a character image.
- Teams already on Google AI Plus / Pro / Ultra subscriptions who want to add AI video to their existing workflow.
FAQ
What is Gemini Omni?
Gemini Omni is Google DeepMind's new multimodal generative model family, starting with video creation and editing. The first member, Gemini Omni Flash, launched on May 19, 2026 and creates or edits video from image, video, text, and supported audio references through conversational instructions.
How is Gemini Omni different from Gemini 3.5 Flash or Pro?
Gemini 3.x Flash and Pro are language and agentic models that output text. Gemini Omni is a generative video model that outputs video and is aimed at creators and editors rather than developer agents and chat assistants. If you're comparing video models specifically, our 14 best AI video generators of 2026 guide covers the leading alternatives.
Where can I use Gemini Omni today?
Inside the Gemini app, Google Flow, YouTube Shorts, and the YouTube Create app. A developer / enterprise API is planned for release within the next few weeks.
How much does Gemini Omni cost?
YouTube Shorts creators can use it for free. Inside the Gemini app and Google Flow it is included with Google AI Plus, Pro, or Ultra subscriptions. API pricing has not been published.
Does Gemini Omni support image and audio output?
Not as standalone output modalities at launch. Gemini Omni Flash currently generates video with audio; Google says image and audio output modalities will be supported in time, with no confirmed date.
Are Gemini Omni videos watermarked?
Yes. Every video generated by Gemini Omni carries Google's invisible SynthID watermark, which lets downstream tools detect that the clip is AI-generated.
Is Gemini Omni available in my country?
Availability varies by geography and by subscription tier. Google states that "features vary by tier and geography" without publishing a country list, so check the Gemini app in your region for current availability.
Can Gemini Omni transfer motion from one video to another character?
Yes. You can supply a reference video for pose and motion plus a separate character image, and Gemini Omni will apply that motion to the character. This is one of the headline features highlighted in Google's launch material.
