Gemini Video
Google's video generator in the Gemini app, now on Gemini Omni Flash: 10-second clips with built-in audio from a text prompt or up to five images
10-40s
Clip, With Extensions
#10
Artificial Analysis Video Arena
$19.99/mo
AI Pro
Overview
Gemini Video is the video generation feature inside Google's Gemini app. You describe a scene in plain language, or upload up to five images, and Gemini returns a 10-second clip with sound generated alongside the picture. You then change it by chatting ("make it night", "move the camera left") instead of starting over. Since Google I/O in May 2026 it runs on Gemini Omni Flash, which replaced Veo 3.1 in the app; the August 2026 Omni 1.1 update added scene extension up to 40 seconds. The same model also powers Google Flow, Google Vids and YouTube Shorts. Inside the Gemini app it needs a Google AI subscription; without one, the free route is short clips in YouTube Shorts and YouTube Create.
The Verdict
Who Should Use Gemini Video?
Best For
- Short social clips, ads and Shorts with sound included
- People already paying for Google AI Plus, Pro or Ultra
- Product shots and photos turned into moving scenes
- Iterating by chat instead of learning a video editor
- Students (free Omni access for a year, limited offer)
Not Ideal For
- Free users who want it inside the Gemini app
- Long or multi-shot stories (10s per clip, 40s max with extensions)
- Volume work: testers hit the limit after 3-5 videos
- Content with celebrities, brands or copyrighted characters (frequent refusals)
- Under-18s, and video-upload editing in the EEA, UK and Switzerland
What's Great
- Audio (dialogue, ambient sound, effects) generated with the video, no separate tool
- Chat-based editing keeps the scene and changes only what you ask; testers call it the standout feature
- "Very professional" output quality for short clips
- Usually renders in under a minute
- Ranks above Veo 3.1 and Kling 3.0 on the Artificial Analysis video arena
- Image-to-video from up to five references; uploaded product details stay intact
- First/last-frame control and 4K upscaling (Omni 1.1, in Flow and the API)
- Free short clips in YouTube Shorts and YouTube Create (18+)
Watch Out For
- No video generation in the Gemini app on the free plan
- Tight, opaque limits: one 10s clip used about 22% of a usage window; Google publishes no video counts
- Frequent refusals around copyright and third-party content
- Backgrounds can warp, clothing colors shift, and frames flicker
- Sometimes ignores negative prompts; signs in the background can be misspelled
- Some jobs reported stuck pending for hours before a retry worked
- Native output is 720p; 1080p and 4K are upscaled
- Visible watermark overlay; in India, South Korea and Vietnam only AI Ultra can turn it off
Pricing
Free
$0
No video in the Gemini app; short clips free in YouTube Shorts and YouTube Create
AI Plus
Local price
Omni video, 2x standard limits, 200 Flow credits/mo
AI Pro
$19.99/mo
Omni video, 4x standard limits, 1,000 Flow credits/mo
AI Ultra
$99.99-$200/mo
5x or 20x Pro limits, 10,000 or 25,000 Flow credits/mo
Gemini API
~$0.10/sec
720p video output, paid tier only, model gemini-omni-1.1-flash
View all features & details
Creation
- Text-to-video, 10 seconds per clip
- Image-to-video from up to five images
- Audio generated with the video; voice references as audio input
- 16:9 or 9:16; a reference image or video sets the ratio
Editing (Omni 1.1, Aug 2026)
- Refine by chat: change lighting, angle, style, add or hide objects
- Scene extension in 10s steps up to 40s (also in the Gemini app)
- First/last-frame interpolation for orbits, zooms and loops (Flow, API)
- 360p drafts: up to 60% faster at a third of the 720p cost (API)
- Edit an uploaded video (10s max) plus up to five images
Limits
- Usage refreshes every 5 hours until a weekly cap
- Video "requires the most usage" of any Gemini feature
- Plus 2x, Pro 4x, Ultra 5x or 20x of Pro limits
- Output 720p native; 1080p and 4K upscaled
Safety & Availability
- SynthID invisible watermark on every video, checkable in Gemini, Chrome and Search
- Visible watermark overlay, controlled under Settings > Media Watermark
- 18+ only; Omni available to subscribers globally
- Video-upload editing not available in the EEA, UK, Switzerland and some US states
Real-World Usage
Hands-On Tests
- Digit (May 2026): quality "very professional for short videos", chat editing the standout, but warping backgrounds and flicker
- Analytics Vidhya (Jun 2026): most clips in under a minute; limit gone after 3-5 videos; refusals the "biggest problem"
- Reddit tester, 3 days (May 2026): talking characters, product shots and photo reels worked; storyboards and long explainers less reliable
Leaderboards
- Artificial Analysis (Oct 2026): Omni 1.1 Flash #10 at 1128 Elo, ahead of Kling 3.0 Pro (#17) and Veo 3.1 (#21), behind Wan 3.0 and Seedance 2.5
- LMArena (Sep 2026, via Gradually): Omni Flash about 1512, ahead of Sora 2 and Runway Gen-4.5
Where Else It Runs
- Google Flow: all Omni 1.1 tools for Plus, Pro and Ultra, paid with Flow credits
- Google Vids: Omni 1.1 at 1080p with duration control for most Workspace business plans (Sep 2026)
- YouTube Shorts and YouTube Create: free for 18+
- Gemini API and Google AI Studio for developers
How It Compares
| Feature | Gemini Video | Kling AI | Runway |
|---|---|---|---|
| Artificial Analysis rank | #10 | #17 (Kling 3.0 Pro) | Not listed |
| Clip length | 10s, extend to 40s | About 10s, extendable | Depends on model |
| Audio with video | Built-in | Yes (3.0) | Depends on model |
| Free option | YouTube Shorts / Create only | Free credits, watermarked | 125 one-time credits |
| Cheapest paid plan | $19.99/mo (AI Pro, US) | Standard plan | $15/mo |
| Chat-based editing | Yes | No | No |
| Best For | Google AI subscribers | Value seekers | Pro filmmakers |
User Reviews
Loading reviews...