Contents
What makes Gemini Omni differentGemini Omni versionsGemini Omni 1.1 Flash specificationsHow creators use Gemini OmniEdit a video with Gemini Omni in RekreatePrompting tips for AI video editsFrequently asked questionsGemini Omni is Google's first any-to-any multimodal model, announced at Google I/O on May 19, 2026. Its first release, Gemini Omni Flash, accepts text, images, audio and video in a single prompt and outputs video — and, crucially, lets you edit and refine that video through natural-language conversation. In August 2026 Google shipped Gemini Omni 1.1 Flash with scene extension, start and end frames, video input for editing and production-ready API access.
Rekreate uses Gemini Omni in its AI Video Editor to change existing short clips with plain-English instructions.
- Announced: May 19, 2026 (Google I/O); Omni 1.1 Flash in August 2026.
- Any-to-any: text, image, audio and video in → video out.
- Conversational editing: describe the change, get an edited clip; iterate.
- Specs (1.1 Flash): 3–10 s output, 360p up to 4K at 24 fps; video input up to 10 s for editing/extension.
- In Rekreate: AI Video Editor for 3–10-second source clips, with history and community gallery.
Edit a video with Gemini Omni
Upload a 3–10-second clip to the AI video editor and describe the change.
Free to try.
What makes Gemini Omni different
Most video models are one-shot generators: prompt in, clip out, and if something is wrong you start over. Gemini Omni is built on Gemini's multimodal reasoning, so it can understand an existing video and apply targeted edits while keeping everything else intact — the subject, lighting, motion and style.
That enables workflows like:
- "Change her jacket to red and make it rain."
- "Replace the coffee cup with our product can."
- "Make this shot golden hour and add a slow push-in."
- "Extend the clip by a few seconds as he walks out of frame."
Google highlights character consistency, accurate physics and SynthID watermarking on every clip. Omni is available in the Gemini app, Google Vids, Flow, Google AI Studio, the Gemini API and the Gemini Enterprise Agent Platform.
Gemini Omni versions
Gemini Omni Flash (preview)
Announced at Google I/O as Google's first any-to-any model. Text and images to video, conversational refinement, character consistency, video extension, upscaling and interpolation.
Gemini Omni Flash in Google Vids and the Gemini app
Rolled into Google Workspace's Vids and the Gemini app; scene extension available to Google AI Plus, Pro and Ultra subscribers.
Gemini Omni 1.1 Flash In Rekreate
Production-ready API release with scene extension, start and end frames, video input for editing and more creative controls. Stable model ID gemini-omni-1.1-flash.
Gemini Omni 1.1 Flash specifications
| Developer | Google DeepMind |
|---|---|
| Model IDs | gemini-omni-1.1-flash (stable), gemini-omni-flash-preview
|
| Inputs | Text, image, audio, video (video up to 10 s for editing and extension) |
| Output | Video, 3–10 s, 360p / 720p / 1080p / 4K, 24 fps |
| Capabilities | Generation, conversational editing, extension, start/end frames, upscaling, interpolation |
| Context window | 1,048,576 tokens |
| Watermark | SynthID |
| In Rekreate | AI Video Editor · source clips 3–10 s |
How creators use Gemini Omni
- Localize a winning ad visually: change the setting, season or props of a proven clip for a different market.
- Product swaps: replace a generic product in stock or generated footage with yours.
- Fix one detail: remove an object, change a color or tweak the lighting without regenerating.
- Style variations for testing: make three versions of the same hook — daylight, neon, cozy interior — and see which performs.
- Polish AI clips from other models: refine a Seedance or Kling generation with a follow-up instruction.
Edit a video with Gemini Omni in Rekreate
Open the AI Video Editor
Go to AI Video Editor in Rekreate.
Upload a 3–10-second clip
Use your own footage or a clip generated with Seedance, Kling or Veo.
Describe the edit
Write what should change and what must stay the same. Use Enhance prompt for clearer instructions.
Review and iterate
Compare results in your history, refine with another instruction, and share your best edits to the community gallery if you like.
Prompting tips for AI video edits
- State what to keep: "Keep the person, framing and motion identical."
- Change one or two things at a time for predictable results.
- Be concrete: "navy blue wool coat" beats "nicer coat".
- Describe light and weather physically: "low warm sun from the left, long shadows."
- Respect rights: don't edit real people into situations without consent or insert trademarks you don't own.
Frequently asked questions
What is Gemini Omni?
Gemini Omni is Google's any-to-any multimodal model, announced May 19, 2026. Gemini Omni Flash takes text, images, audio and video as input and generates or edits video through natural-language conversation.
What is new in Gemini Omni 1.1 Flash?
Released in August 2026, Omni 1.1 Flash adds scene extension, start and end frames, video input for editing and production-ready API availability in Google AI Studio, Flow and the Gemini Enterprise Agent Platform.
How long are Gemini Omni videos?
Gemini Omni 1.1 Flash outputs 3–10-second videos, and accepts videos up to 10 seconds for editing and extension.
Is Gemini Omni the same as Veo?
No. Veo is Google's dedicated video generation family (latest Veo 3.1). Gemini Omni is a separate any-to-any model focused on conversational generation and editing.
Can I use Gemini Omni to edit my TikToks?
Yes, for short clips. In Rekreate's AI Video Editor you can upload 3–10-second clips and edit them with text instructions.
Does Gemini Omni add a watermark?
Every Gemini Omni clip includes Google's invisible SynthID watermark.
Create your next viral video with Rekreate
Find a proven short-form video, recreate it with your script, avatar and voice, and generate scenes with Seedance, Kling and Veo in one workspace.
Start free trial →