Gemini Omni —
Speak it. See it. Share it.

How to use Gemini Omni

Gemini Omni is dead simple to use — it's the videographer friend you can ask for things in plain English, and a short clip comes back. No timelines, no plug-ins, no render bars to stare at.

The easiest place to start is the Gemini app, on your phone or in your browser. Open a new chat, pick Gemini Omni Flash from the model dropdown, and type what you want to see — say, "a 10-second clip of a violinist on a quiet rooftop at golden hour." About a minute later, the video shows up in the chat. That's the whole thing.

If you'd rather work visually, open Google Flow at labs.google/flow. Flow is Omni's visual workspace — drop in up to five reference photos, line up shots like a storyboard, and tweak details by talking ("add rain, keep her expression"). Put another way: editing, without an editing app.

If you write code and want Omni inside your own product, that's the Gemini Omni API — find it in Google AI Studio. Same model, just plugged into your code instead of a chat window. Not a developer? Skip this paragraph.

• Price. Any Google AI plan (Plus, Pro, Ultra) covers it. In YouTube Shorts and YouTube Create, Omni is free starting this week.
• Length. Up to 10 seconds per clip — roughly a TikTok intro. For longer pieces, stack a few short clips in Google Flow.
• Where it works. Anywhere the Gemini app is available. A handful of features (avatars, editing existing footage) open country by country.

One last tip — worth more than any setting: talk to it the way you'd talk to a friend, not the way you'd type into a search box. "Show me a busy ramen shop at midnight, steam off the pots, neon outside the window" gets you somewhere much better than "ramen night neon video."

Where Gemini Omni shows up in Google's tools

Gemini Omni isn't tied to one app — it shows up across several Google products, each doing a different job. Knowing which one lives where saves a lot of trial and error.

• Google Flow is the studio. If "make a video" is the goal, this is the room. Timeline, reference photos, prompt history — all in one place. People search both google flow omni and google omni flow; same product, two folk names.
• Google Vids is the office tool. It lives inside Google Workspace next to Docs, Slides, and Meet. More internal explainers and team how-tos than cinema — think PowerPoint that moves.
• Google Pics is the image-side sibling, and home to Nano Banana, Google's image model. The handy bit: a still you generate in Pics can be sent straight into Omni and start moving — no re-upload, no style drift.
• Google AI Studio and Google Labs are the workbenches. AI Studio is mostly for developers — Omni and Gemini 3.5 sit side-by-side for testing. Labs is where Google puts early experiments out for anyone to play with — Flow itself started life in Labs.
• Google Spark and Gemini Spark are the personal assistants. Tell Spark "make a little birthday video for my dad" and Spark quietly calls Omni behind the scenes — the way you'd ring up a friend who happens to be good at editing.

What ties them together is the same brain. The video Flow renders, the image Pics paints, the avatar in your meeting — all the same Gemini Omni model under the hood, just dressed for different work. The reference photo you saved in one app last week is usable in another app today, no setup required.

Gemini Omni next to Veo 3, Sora 2, Runway, HeyGen, and Higgsfield

Picking an AI video tool is a bit like picking a kitchen knife — they all cut, but each has something it's best at. Here's a quick map to save you a lot of exploring.

• Veo 3 is Google's previous video model — Omni now handles the day-to-day slot in the Gemini app. Veo 3 isn't gone (enterprise customers still use it through Vertex AI), it just isn't the default anymore. If you were using veo3 before, opening the app today already puts you on Omni.
• Sora 2 is OpenAI's closest match. Same broad idea — type, get a 10-to-20-second clip — and physics (water, hair, fabric) feels solid. Day to day, the difference is which world you live in: Sora 2 sits inside ChatGPT's world; Omni sits inside Google's (Docs, Photos, YouTube). Pick by where your stuff already lives.
• Runway is the short-film veteran's pick. Think chef's knife: more knobs, more precision, more learning curve. If you already work in After Effects or Premiere, Runway feels at home. Otherwise, the button count can scare you off.
• HeyGen does one thing very well — a person on camera reading a script in their own voice. Sales videos, corporate training, language tutorials. Omni's avatar handles the short version of this; for anything past a couple of minutes, HeyGen is the better fit.
• Higgsfield is the cinematographer. If you can already see the shot in your head ("slow push-in, then crane up over her shoulder"), Higgsfield gives you the controls to spell that out. Most people don't need to go that deep; the ones who do, swear by it.
• Grok is the model inside X (formerly Twitter). Its images and short clips are optimized for posts — paste straight into a tweet and go. Social-first, low-friction, not really the place for serious creative work.

In one line: Omni is for talking your way to a clip; Sora 2 is its twin in a different ecosystem; Runway is for precision editing; HeyGen is for on-camera scripts; Higgsfield is for cinematography; Grok is for posts. Most people end up using two — Omni for the first cut, plus a specialist tool for whatever specialty the project needs.