Gemini Omni AI Video Generator

Turn a prompt or your images into a video with sound with Google’s Omni 1.2 Flash: realistic motion, frames you choose and up to 6 reference images, in up to 4K. Online in your browser, no Google subscription needed.

Try Omni 1.2 Flash online

Model

0/20,000

Aspect ratio

Quality

5s

Audio

Native audio, always on

5s · 360p · with audio · No watermark

Example · Gemini Omni
  • Gemini Omni 1.2 Flash
  • Text to Video
  • Image to Video
  • Reference to Video

Google’s Gemini Omni launch film. Write a prompt or add your images to make your own.

What is Gemini Omni?

Gemini Omni is Google’s family of multimodal video models, introduced at Google I/O in May 2026. Omni Flash came first, 1.1 Flash followed in August 2026 and 1.2 Flash, the version this page opens on, in November 2026. It makes a video with sound from a text prompt, from a first and last frame, or from reference images. On Melies AI you can use it as an online Gemini Omni video generator, straight from your browser: no Google AI subscription, no API key and no GPU of your own.

Google describes Omni as a model with an intuitive grasp of forces like gravity, kinetic energy and fluid dynamics, so movement looks real: things fall, splash and collide the way they should. It follows camera directions closely, from a push-in or a dolly zoom to one continuous shot, and keeps a scene consistent with the images you give it.

Omni 1.2 Flash makes any length from 3 to 15 seconds (up to 10 with a last frame), in 16:9 or 9:16, with speech, music and sound effects made with the picture. Draft in 360p, then render the take you like in 720p, 1080p or 4K. The older 1.1 Flash stays in the model menu for fixed 4, 6, 8 or 10 seconds clips and up to 7 reference images. Prompts can run up to 20,000 characters, and every video downloads as an MP4 with no visible watermark.

Clip length
3–15s
Top quality
4K
Reference images
6
Quality levels
4
Aspect ratios
2
Credits · 5 s at 720p
15

Gemini Omni 1.2 and 1.1 Flash by Google · run through a third-party API · not affiliated with Google · Last updated

Omni Flash features

What Google shows Gemini Omni doing in its launch film, and how to get the same results here. The generator above is already set to Omni 1.2 Flash.

@Image1The same person in a new world
01

One person, any world

In Google’s film, one face becomes a clay-animation driver, a studio portrait and a swimmer under water, and stays the same person in all three. Upload photos of your character as reference images (up to 6), call them @Image1 and so on, and describe the world you want them in.

@Image1 + @Image2Both in one scene, with your title
02

Two characters in one scene, title included

Google’s “girls trip” demo puts two friends in the same car, then ends on a 3D “GIRLS TRIP” title standing in the desert. Add one image per character, say what they do together, and write any title or sign you want in the shot in quotes. Keep on-screen text short.

PromptThe same scene in the style you named
03

Any style, from collage to chrome

In the film, a guitarist on the clouds turns into a paper collage, a risograph print and a chrome statue. Gemini Omni moves between live action, illustration, animation and 3D in one model: name the look in your prompt, such as “paper collage”, “risograph print” or “polished chrome sculpture”.

360pThe same prompt in 1080p or 4K
04

Draft in 360p, finish in 4K

Google’s side-by-side shows the same shot at 720p and at 4K, where fish scales and fur get sharp. Omni 1.2 Flash renders in 360p, 720p, 1080p, 4K: try a prompt in 360p, the cheapest draft at 1 credit a second, then render the version you like in 1080p, or in 4K when it goes on a big screen.

Omni 1.2 Flash vs 1.1 Flash vs Seedance 2.5 vs Grok Imagine 1.5

Four of the models in the same generator. Omni 1.2 Flash makes up to 15 seconds with sound and goes up to 4K; 1.1 Flash takes one more reference image in fixed lengths; Seedance 2.5 makes the longest clips; Grok Imagine 1.5 adds loops and keyframes.

ModelGemini Omni 1.2 FlashGemini Omni 1.1 FlashSeedance 2.5Grok Imagine 1.5
Clip length3–15 s4, 6, 8, 10 s4–30 s1–15 s
Quality360p, 720p, 1080p, 4K360p, 720p, 1080p, 4K480p, 720p, 1080p480p, 720p, 1080p
Aspect ratios2278
Reference imagesUp to 6Up to 7Up to 30Up to 1
First and last frameYes, up to 10 sYesYesYes
MakerGoogleGoogleByteDancexAI
Price at 720p3 credits / s4 credits / s15 credits / s2 credits / s
Price at 1080p3 credits / s4 credits / s36 credits / s2 credits / s
Price at 4K5 credits / s9 credits / s——
Best forClips up to 15 s with sound, from 360p drafts to 4KFixed 4 to 10 s clips with up to 7 reference imagesClips up to 30 s, many references, final rendersLifelike motion and speech, loops and keyframes

Need clips longer than 15 seconds? See Seedance 2.5

How to use Omni Flash

  1. 01

    Choose a mode

    Text to video for a scene from scratch, image to video to start on your picture (and end on another), or reference to video to build a scene around your characters and products. Omni 1.2 Flash is already selected.

  2. 02

    Write your prompt

    Describe the shot, the action, the light, the camera move and the sound. Google’s tip: the more detail you add, the more control you get. Call your images @Image1, @Image2 and so on.

  3. 03

    Set length, format and quality

    Pick any length from 3 to 15 seconds (up to 10 with a last frame), 16:9 or 9:16, and 360p to 4K. The price is on the button: a 10-second 1080p clip is 30 credits.

  4. 04

    Generate and download

    Press Generate. The video renders while you keep browsing, then download the MP4.

Omni Flash pricing

Pay per second of video; all three modes cost the same. On 1.2 Flash, 360p is the cheapest draft and 720p and 1080p share one price; 1.1 Flash has one price from 360p to 1080p. Packs from $9.90, no subscription, failed videos refunded.

Quality5 s10 s15 s
Gemini Omni 1.2 Flash · 3 to 15 s · sound · up to 4K
360p · Draft5 credits10 credits15 credits
720p · HD15 credits30 credits45 credits
1080p · Full HD15 credits30 credits45 credits
4K · Ultra HD25 credits50 credits75 credits
Gemini Omni 1.1 Flash · 4 to 10 s · up to 4K
360p · Draft—40 credits—
720p · HD—40 credits—
1080p · Full HD—40 credits—
4K · Ultra HD—90 credits—

Google Omni video FAQ

Something else? Email support@meliesai.com.

Gemini Omni is Google’s family of multimodal AI video models. It makes video from text, images and video, with an understanding of real-world physics and camera language. The newest, Omni 1.2 Flash, makes 3- to 15-second clips with sound in up to 4K. On Melies AI you can use it in your browser from a prompt, a first and last frame or up to 6 reference images.

Google introduced its Omni video model at Google I/O in May 2026, with Omni Flash as the first version. Omni 1.1 Flash followed in August 2026, and Google published Omni 1.2 Flash in November 2026. Melies AI runs both 1.2 and 1.1 Flash, and the generator opens on 1.2.

In Google’s own apps Gemini Omni needs a Google AI subscription, with features that vary by plan and country. Here there is no subscription: you pay per video with credits. Packs start at $9.90 for 100 credits, enough for 6 clips of 5 seconds in 1080p on Omni 1.2 Flash, and failed videos are refunded automatically.

Omni 1.2 Flash is priced per second: 1 credit at 360p, 3 at 720p or 1080p and 5 at 4K, so a 5-second 1080p clip costs 15 credits and a 15-second 4K clip 75. Omni 1.1 Flash costs 4 credits per second from 360p to 1080p and 9 at 4K. Text, image and reference to video cost the same.

Omni 1.2 Flash makes any length from 3 to 15 seconds (up to 10 with a last frame); 1.1 Flash makes 4, 6, 8 or 10 seconds. Pick the length before you generate. For longer clips, Seedance 2.5 and Wan 3.0 in the same model menu go up to 30 seconds.

Yes. Omni 1.2 Flash renders in 360p, 720p, 1080p, 4K. 4K costs 5 credits per second against 1 at 360p and 3 at 720p or 1080p, so draft in 360p and render your final take in 4K.

16:9 or 9:16: landscape for YouTube and the web, vertical for TikTok, Reels and Shorts. In reference to video, Omni 1.2 Flash also has Auto. In image to video, use a first frame in the same shape as the ratio you pick.

Pick the Reference tab and add up to 6 images on 1.2 Flash (7 on 1.1 Flash), for example a character, an outfit, a product and a location. They are numbered in the order you add them; refer to them in the prompt as @Image1, @Image2 and so on. Reference images and a first frame cannot be combined in one video.

Gemini Omni 1.2 Flash is the one for 4K and for motion with real-world physics, in 3- to 15-second clips with sound, at 3 credits / s in 1080p. Seedance 2.5 makes clips up to 30 seconds, takes up to 30 reference images and has seven aspect ratios, at 36 credits / s in 1080p. Both are in the same model menu, so you can try a prompt on each.

Grok Imagine 1.5 by xAI makes 1 to 15 seconds in up to 1080p for less per second at 720p and 1080p, with loops and keyframes. Gemini Omni 1.2 Flash by Google goes up to 4K, takes up to 6 reference images instead of 1, and is built for physically believable motion.

No. Veo is Google’s earlier video model line, built for text and image to video. Gemini Omni is a newer, multimodal family: in Google’s own apps it also edits and extends videos through conversation. Veo is not on this site; Gemini Omni 1.2 and 1.1 Flash are.

Yes. Google published Gemini Omni 1.2 Flash in November 2026, and the generator on this page opens on it. Compared with 1.1 Flash it makes any whole length from 3 to 15 seconds instead of four fixed ones, generates speech, music and sound effects with the picture, costs less per second and adds Auto for reference to video; it takes 6 reference images instead of 7.

Not on Melies AI. In Google’s apps Omni can edit a video through conversation, extend a scene and take video or voice references. Here Omni 1.2 and 1.1 Flash make new videos from text, a first and last frame, or reference images. For editing an uploaded video, use Kling 4.0’s Edit tab in the same menu.

There is no visible watermark or logo on videos made on Melies AI. Google embeds SynthID, an invisible digital watermark that identifies AI-generated content, in what its models make; it does not change how the video looks, and we do not remove it.

Google’s advice is that the more detail you add, the more control you have. Cover the framing and camera move, the style, the lighting, the location, the action and the sound, in that order or any other. Use film terms such as “push in”, “dolly zoom” or “one continuous shot”, and add what you do not want, such as “Don’t add text.”

Open the generator at the top of this page, where Omni 1.2 Flash is already selected. Pick a mode, write a prompt or upload images, choose length, aspect ratio and quality, and press Generate. The finished MP4 appears under “My videos”.

Yes. Videos you generate are yours to use, including in ads, client work and monetized posts, as long as your prompt and images do not infringe anyone else’s rights and follow our content standards.

No. Melies AI is a web app: you use Gemini Omni in the browser and need no API key or code. Developers who want to call the Gemini Omni API from their own software can use Google AI Studio, Vertex AI or an API provider.

No. Melies AI is an independent platform that gives you access to Google’s Omni 1.2 and 1.1 Flash through a third-party API provider. It is not affiliated with, endorsed by or sponsored by Google. Finished videos stay under “My videos” for 14 days.

Make your first Omni Flash video

Write a prompt or upload an image, pick a length and render it with sound in up to 4K.

Start creating