Yes. As of October 2026, Gemini makes video clips with Gemini Omni, the model that replaced Veo 3.1 in the Gemini app. You need a Google AI Plus, Pro or Ultra plan and must be 18 or older. You describe a scene or add up to five photos, and Gemini returns a clip of up to 10 seconds with native audio that you can edit by chatting or extend. Every clip carries an invisible SynthID watermark. Gemini can also analyze videos you upload and YouTube videos. A finished Short still needs a script, voiceover, captions and an edit.
This guide covers the current facts (model, plans, clip length, Flow, watermarks, video analysis) and then the practical part: why a 10-second clip isn't a video yet, a workflow from script to published video, Gemini vs ChatGPT vs Claude, disclosure rules and common mistakes. Everything is stated as of October 2026.
Related guides: can Claude make videos, Sora 2 vs Veo 3, and the best AI video generators in 2026.
Which Model Does Gemini Use for Video in 2026?
Gemini's video generator is Gemini Omni. Google introduced the Omni family at Google I/O on May 19, 2026, describing Gemini Omni Flash as a model that can create anything from any input, starting with video, and rolled it out to Google AI Plus, Pro and Ultra subscribers in the Gemini app and Google Flow (Google blog). Google says Omni replaces Veo 3.1 in the Gemini app, and Gemini Apps Help names it as the model behind Create video (Gemini Apps Help).
Omni blends text, images, video and voice references into one generation, and you can keep editing a clip across several turns: replace objects, change camera angles or modify the scene. Google's Omni 1.1 Flash update on August 27, 2026 brought scene extension to the Gemini app for Plus, Pro and Ultra subscribers.
Veo hasn't disappeared. Google Flow still offers Veo 3.1 Lite, Fast and Quality next to Gemini Omni Flash 1.1. A guide saying “Gemini uses Veo 3” was right in 2025 but not for today's Gemini app; our Sora 2 vs Veo 3 comparison covers that earlier generation.
Is Gemini Video Generation Free? Plans and Limits
Not in the Gemini app. Video generation needs a Google AI plan on a personal account (or a qualifying Workspace license) and isn't available under 18; Google's plan table includes it in Plus, Pro and Ultra only (Gemini Apps limits).
US prices on Google's subscriptions page in October 2026: Google AI Plus is $4.99 a month with 200 Google Flow credits, Google AI Pro is $19.99 with 1,000 Flow credits, and Google AI Ultra is $99.99 (5x Pro's usage) or $199.99 (20x), with 10,000 or 25,000 Flow credits. Prices and availability vary by country.
There's no fixed “videos per day” number anymore. Gemini Apps use compute-based limits that refresh every five hours up to a weekly cap; Plus gets 2x the standard limit, Pro 4x, and Ultra 5x or 20x Pro's, and media generation uses more allowance than text. Articles quoting two, three or five Veo videos a day describe the 2025 system. Check Settings, then Usage Limits.
Free routes exist outside the Gemini app. Google made Omni available at no cost in YouTube Shorts Remix and the YouTube Create app for users 18 and over, and in September 2026 brought Omni to Google Vids, where anyone with a Google account can generate 1080p clips at no cost.
How Long Can Gemini Videos Be?
A single Gemini Omni generation is up to 10 seconds. In Flow, Omni Flash 1.1 offers 4, 6, 8 or 10-second clips, while Veo 3.1 offers 4, 6 or 8 seconds (Flow models and features). The familiar 8-second figure comes from Veo.
Scene extension continues a clip in 10-second steps, and Google's Omni 1.1 Flash announcement caps an extended clip at 40 seconds in total. Anything longer is an edit: several clips joined on a timeline.
On resolution, Flow lists Omni at a standard 720p plus a 360p draft mode at half the credit cost, Google Vids generates 1080p, and Google's developer API offers 1080p and 4K outputs. The Gemini app defaults text prompts to landscape, so choose portrait before generating a clip for Shorts, Reels or TikTok; an uploaded reference image or video sets the aspect ratio to match.
How to Make a Video with Gemini, Step by Step
Step 1: open gemini.google.com, open the sidebar and choose Create video (on mobile it's in the menu). Optionally pick a style template.
Step 2: write the prompt like a shot description: subject, action, setting, camera movement, lighting, style and sound. “Close-up of hands kneading dough on a floured wooden table, slow push-in, warm morning light, soft kitchen ambience, no music” beats “baking video.”
Step 3: add references: one video and up to five images, such as a product photo or character sheet. If you've set up a Gemini avatar, typing @ plus your Google username adds your own likeness and voice. Only upload media you have rights to.
Step 4: set the aspect ratio and submit. Generation takes a few minutes; ask for audio in the prompt if you want dialogue or ambience.
Step 5: edit by conversation: remove an object, swap the background, change the angle or extend the scene. Editing an uploaded video isn't available in the EEA, Switzerland, the UK and some US states.
Step 6: under the video, choose Share, then Download video or Share on YouTube.
What Is Google Flow, and When Should You Use It?
Google Flow is Google's AI filmmaking studio for clips, scenes and stories with consistent characters. You must be 18 or older, live in a supported region and subscribe to Google AI Plus, Pro or Ultra; qualifying Workspace plans get 50 daily credits. Flow runs on credits rather than the Gemini app's limits, and failed generations aren't charged.
Flow adds controls the Gemini app hides: a choice of Veo 3.1 or Omni Flash 1.1, start and end frames, ingredients (reference images for characters, products and locations), custom voices, a Flow Agent that runs multi-step jobs such as several lighting variations at once, and scene building from clips. Use the Gemini app for quick single clips and photo-to-video; use Flow when shots must match, as in a short story or an ad with a recurring product (see how to make video ads with AI).
Watermarks: SynthID, C2PA and the Visible Sparkle
Every video made in Gemini or Flow carries an invisible SynthID watermark, and Google says it also attaches C2PA Content Credentials metadata; the Gemini app can check whether Google AI made a file.
The visible sparkle in the corner is now optional in many places. In August 2026 Google added a setting to hide it on new content in Gemini and Flow; in Flow it's the Visible watermarking toggle under your profile picture. Flow applies the visible watermark automatically for users in India, South Korea and Vietnam (Google Flow Help), and Google has said the toggle won't appear where law requires a visible mark. Hiding the sparkle doesn't make a clip undetectable: YouTube and TikTok read Content Credentials and may label uploads automatically.
Can Gemini Watch and Analyze Videos?
Yes, and for many creators this is the more useful half. You can upload up to 10 files per prompt, each video up to 2 GB, with a total video length of 5 minutes, or 1 hour on Google AI Pro and Ultra (Gemini Apps Help). Gemini can also find YouTube videos and answer questions about them when you're signed in with Keep Activity on.
Useful jobs: ask why a competitor's top Short holds attention in its first three seconds, ask where your rough cut drags, or get a timestamped summary of a long video to plan clips.
Why a 10-Second Clip Isn't a Finished Video
A generated clip is a shot, not a video. A Short that performs has a hook in the first second or two, a reason to keep watching, a payoff and a voice, with captions for people watching on mute. Even a 40-second extended clip is one continuous scene, while a 45-second Short usually works best with a visual change every two to four seconds.
Generated clips also drift between generations, garble on-screen text and spend usage on every retry. YouTube's monetization policies exclude repetitive, mass-produced content, so stitched clips without an original script are a weak channel strategy; our guide to YouTube's AI content monetization policy explains what reviewers look for. Treat Gemini as a footage source inside a normal production workflow.
Workflow: Turn Gemini Clips into a Full Short or YouTube Video
Step 1, script: write the narration first; a 45-second Short is about 110 words at a natural pace. Our YouTube script guide and hooks guide show the structure.
Step 2, shot list: mark each line's visual and decide which shots truly need generation and which can be stock footage, screen recordings or phone footage. Spend generation on the hook and one or two key beats.
Step 3, generate: make clips in the final aspect ratio (9:16 for Shorts, 16:9 for YouTube), reusing the same references and style words so shots match.
Step 4, voiceover: record yourself or use a text-to-speech voice, and keep one voice across the channel; our AI voiceover tools comparison covers options. Mute or duck the clips' generated audio under the narration.
Step 5, edit and sequence: lay the voiceover on a timeline, cut clips to it, and fill gaps with b-roll and a quiet music bed.
Step 6, captions: burn in captions for Shorts and upload an accurate caption file for long-form; a subtitle SRT generator makes the file.
Step 7, packaging: write the title and thumbnail together, set the AI disclosure where it applies, and check the YouTube Shorts length and size requirements. For long-form, our guide to long YouTube videos with AI covers chapters and retention.
The Fast Way: Full Narrated Videos with Vidpal
If you want a finished narrated video rather than a cinematic shot, stitching clips is the slow route. Vidpal Studio builds the whole video from a topic, your own script or an article URL. You review and edit the script first, then Studio adds a voiceover from its voice library, visuals from stock footage or AI-generated images in 20 styles shown with camera motion, and captions in one of 75 presets. A 9:16 Short costs 1 credit; a 16:9 Long Video of 1 to 10 minutes costs 5.
Vidpal doesn't use Veo or Gemini Omni; Studio's main visuals are stock footage or AI stills. For a generated moment, Studio's Hook Intro can open with a short Text → Video or Image → Video clip, a library clip or a text reveal, and you can regenerate any section's image or swap in stock footage.
If you've already cut Gemini clips together with your narration, upload the edit to Vidpal's AI Editing for automatic captions, silence and filler-word removal, auto zooms and b-roll. Since Gemini defaults to landscape, Smart Reframe (paid plans) turns a 16:9 edit into a vertical Short; see our landscape-to-vertical guide. Finished videos can go out through the scheduler to YouTube, Instagram, TikTok and LinkedIn on Pro or Business, and exports need a paid plan starting at $19 a month (pricing).
Gemini vs ChatGPT vs Claude for Video
Gemini: the only one of the three with built-in video generation in October 2026. It generates clips with audio on paid plans, edits them by conversation, offers Flow for multi-shot work, and analyzes videos.
ChatGPT: OpenAI's help center says the Sora app and website were discontinued on April 26, 2026 and the Sora API on September 24, 2026, so ChatGPT users no longer have OpenAI's dedicated video generator. ChatGPT still writes scripts, creates images and can operate video tools through connectors.
Claude: no video generation and no video file input. Claude's value is writing, code-based video through Claude Code, the Claude Motion beta for teams, and MCP connectors that run tools like Vidpal; see can Claude make videos.
A common split: Gemini for shots and analysis, Claude or ChatGPT for scripts, and a video tool for the narrated final cut.
AI Disclosure Rules on YouTube, TikTok and Instagram
YouTube: disclose meaningfully altered or synthetic content that looks realistic, such as a real person saying something they didn't, altered footage of real events or places, or realistic scenes that never happened. Clearly unrealistic content, minor aesthetic edits and AI help with scripts, titles or thumbnails don't need it. YouTube may label content made with its own AI tools or carrying C2PA metadata automatically, and says disclosure won't limit a video's audience or eligibility to earn money (YouTube Help).
TikTok: realistic AI-generated content needs TikTok's AI-generated content label, and TikTok can apply it automatically when a file carries Content Credentials. Instagram: Meta's “AI info” label applies to photorealistic video or realistic audio that was digitally created or altered, and Meta may add it from industry metadata.
A photorealistic Gemini clip of a scene that never happened needs the label on all three; a stylized clip usually doesn't. Our AI content labels guide covers each toggle.
Common Mistakes with Gemini Video
Expecting a full video from one prompt: a generation is up to 10 seconds, so plan shots, not videos.
Generating in the wrong aspect ratio: text prompts default to landscape; set portrait first for Shorts, Reels and TikTok.
Spending usage on shots stock footage could cover: save generation for moments that can't be filmed or found.
Letting characters drift: reuse the same references and style words, or build multi-shot projects in Flow.
Assuming no sparkle means no label: SynthID and Content Credentials stay in the file.
Key Takeaways
Takeaway one: Gemini makes videos with Gemini Omni, which replaced Veo 3.1 in the Gemini app; Flow still offers Veo 3.1 alongside Omni.
Takeaway two: the Gemini app needs a paid Google AI plan; free Omni access exists in YouTube Shorts Remix, YouTube Create and Google Vids.
Takeaway three: a generation is up to 10 seconds, extendable in 10-second steps to about 40 seconds, and every clip carries SynthID.
Takeaway four: Gemini also analyzes uploaded and YouTube videos.
Takeaway five: a finished video still needs a script, voiceover, captions and an edit, which Vidpal Studio handles end to end.
Frequently Asked Questions
Can Gemini make videos for free? Not in the Gemini app, which requires a Google AI Plus, Pro or Ultra plan and users 18 or older. Gemini Omni is free in YouTube Shorts Remix and the YouTube Create app for adults, and in Google Vids for anyone with a Google account.
Does Gemini still use Veo? Not in the Gemini app. Google says Gemini Omni replaces Veo 3.1 there, while Google Flow still offers Veo 3.1 Lite, Fast and Quality alongside Gemini Omni Flash 1.1.
How long can a Gemini video be? A single Gemini Omni generation is up to 10 seconds. Scene extension adds 10-second steps up to 40 seconds in total, per Google's Omni 1.1 Flash announcement; longer videos are edited from several clips.
How many videos can I make with Gemini per day? Google no longer publishes a fixed number. Compute-based limits refresh every five hours up to a weekly cap, video uses more allowance than text, and higher plans get more.
Do Gemini videos have a watermark? Yes. Every clip includes an invisible SynthID watermark plus Content Credentials. The visible sparkle can be hidden in many countries, but Flow applies it automatically in India, South Korea and Vietnam.
Can Gemini analyze a YouTube video? Yes. Gemini can find YouTube videos and answer questions about them when you're signed in with Keep Activity on. You can also upload your own videos, up to 5 minutes in total without Google AI Pro or Ultra, or 1 hour with them.
Can I post Gemini videos on YouTube Shorts and TikTok? Yes. Download clips or share them to YouTube from Gemini, and disclose realistic AI-generated footage on YouTube, TikTok and Instagram, which may also label it automatically from metadata.
Does Vidpal use Gemini or Veo? No. Vidpal Studio builds narrated videos from a topic, script or URL using stock footage or AI images with camera motion. You can upload your own edit of Gemini clips to Vidpal's AI Editing for captions and cleanup.
Make Your First Full Video
Of the major AI assistants, Gemini is the one that generates video in 2026. Use it for shots that can't be filmed, then build the video around a script, a voice, captions and an edit.
To go straight from an idea or article to a finished narrated video, open Vidpal Studio and choose Short or Long Video. You can create a free account to explore, and read how to make YouTube Shorts with AI for the full Shorts playbook.
