Creation settingsGemini Omni 1.2 Flash

Aspect ratio

Resolution

A mustard-yellow car rounds a grassy coastal road above the sea

Gemini Omni 1.2 Flash AI video generator

Gemini Omni 1.2 Flash is a video generation and editing model that works from a written prompt, still images, or existing footage. It can make clips up to 15 seconds long with synchronized sound. Start with a scene description, give the shot an opening frame, or supply references for details you want to carry into the footage.

Open video workspace
A swimmer crosses a desert motel pool framed by terracotta walls and distant mountains

Gemini Omni 1.2 Flash features

Plan a shot around the material you already have. A written scene gives the camera room to move, a first frame sets the composition, and a reference helps carry a specific detail through the action.

A commuter train moves along a steel bridge in the morning haze

Direct a clip from text

Set a clip length from 3 to 15 seconds and describe what happens in that time. A short clip might follow a train as it crosses a bridge. A longer one can begin inside the carriage, cut to the bridge, and hold on the train as it leaves the frame. You can specify camera movement or ask for a continuous take.

Open video workspace
Two frames show the same hiker at different points on the trail to a mountain hut

Animate an image or connect two frames

Use a still as the opening frame when the subject and composition are already set. In First and last frames mode, a second image gives the shot a destination. For example, start with a hiker on a trail and end at the door of the same hut. Clips made with both frames run from 3 to 10 seconds.

Open video workspace
A mint-green radio with two knobs sits on a sunny tiled balcony

Carry details through a moving shot

Reference images can guide the look of a character, product, or location as the camera moves. A short motion reference can show the path or timing of the movement. For a product clip, a front view can establish the controls and a side view can show the silhouette the model needs to follow.

Open video workspace
Two people talking over a cup at a daylight market stall

Generate sound with the picture

Dialogue, ambience, and sound effects can be generated along with the video. In a two-person scene, the model can distinguish the speakers and synchronize dialogue with their mouth movements. The same approach works for small sounds tied to visible action, such as a cup meeting a counter.

Open video workspace
Two views of the same bakery entrance show a bicycle with a changed bag and warmer light

Edit or extend existing footage

Provide a clip and describe the change you want, such as replacing a bicycle bag or warming the light. Edit video works with a source clip of up to 15 seconds and keeps its duration. Extend video adds 3 to 10 seconds to the end of the shot when the action needs more time.

Open video workspace

How to use Gemini Omni 1.2 Flash

The workspace changes its inputs and output settings to match the mode you select.

  1. Choose a mode

    Start with Text to video for a new scene, Image to video for an opening still, or First and last frames for a shot with a defined ending. Use Reference to video when the media describes a subject or movement instead of a particular frame.

  2. Describe the shot and add media

    Enter a prompt and add the files the mode asks for. For an edit, select the source clip and describe the specific change. For an upscale, select the source clip and choose an output resolution.

  3. Set the output

    Choose duration, aspect ratio, and resolution where the mode offers them. Use 16:9 for landscape framing or 9:16 for portrait framing. The preview workspace lets you check these settings before generation becomes available.

Compare video and image models

Choose by the output you need. A still image can also serve as the opening frame or visual reference for a video.

CapabilityGemini Omni 1.2 FlashNano Banana 2.1GPT Image 2.5Nano Banana Pro
OutputVideo with soundStill imageStill imageStill image
Start fromText, images, videoText or imagesText or imagesText or images
Edit sourceVideo clipStill imageStill imageStill image
Example useProduct motionReference-based visualsPoster layoutsProduct mockups

Explore more AI image models

Make a still for its own use or bring it into the video workspace as a starting frame.

Cobalt ceramic pitcher and sand-colored cup on a terracotta tile

Nano Banana 2.1

Develop a product still or a character reference, then refine the image before using it to guide a shot.

Cream Blue Hour jazz poster with an indigo saxophone on an olive wall

GPT Image 2.5

Build a poster, title card, or other design where the placement of text matters.

Peach-colored cosmetic bottles arranged on stepped concrete plinths

Nano Banana Pro

Make a product mockup with a specific arrangement, backdrop, and lighting.

Explore more AI tools

Go directly to an existing workspace for the next part of your project.

Illustrated hillside town with yellow roofs and blue shadows

AI image generator

Start with a written scene and make a still to use as a reference or storyboard frame.

Two versions of a red chair in a room with different wall finishes

AI image editing

Bring an existing image into the creator and change a subject, a background, or another detail.

A roller skater turns across a sunlit concrete skate park

AI video generator

Open the live video studio to generate a clip with a model already available there.

Frequently asked questions

What is Gemini Omni 1.2 Flash?

It is a model for generating and editing short video with synchronized sound. It can start with a text prompt, use still images to guide the picture, or take a video clip as the source for an edit.

What is Gemini Omni 1.2 Flash good at?

It suits short sequences where motion and continuity matter. Examples include following a product as it turns, animating a shot between two planned frames, or creating a dialogue scene with two speakers.

Can Gemini Omni 1.2 Flash edit a video I already have?

The model supports prompt-based edits to clips up to 15 seconds long. Select Edit video in the preview workspace to attach a clip and describe a change to the scene, such as a different background or time of day.

Can it render text inside the video?

The model can carry text from a visual reference into the footage and can generate short on-screen labels. Check spelling and legibility as the camera moves, especially when a label turns away from the camera or gets smaller in the frame.

How can I try it on this page?

Choose a mode, write a prompt where required, and select any media and output settings. This workspace is a preview, so generation is currently unavailable. Select a thumbnail to view a different example frame.

Which modes can I preview?

The workspace includes Text to video, Image to video, First and last frames, Reference to video, Edit video, Extend video, and Upscale video. The mode you pick determines whether you need an image, a video, or a prompt.

Plan your next shot on Nano Banana 2

Use the Gemini Omni 1.2 Flash workspace to plan a clip around your own material. Bring an opening frame or a reference from a Nano Banana 2 image tool, then choose the mode, framing, and duration that suit the shot.

Open video workspace
Storyboard prints, a camera lens and a yellow mug on an olive studio desk in afternoon light