All updates
New Model

New model: Gemini Omni Video — 4K with sound

Video up to 4K with a native audio track, character consistency from references and multimodal input.

A film strip and ribbons of sound waves made of light

Gemini Omni Video is an advanced multimodal video model that generates clips at up to 4K resolution and adds a native audio track straight from the model, with no separate voice-over step. It keeps a character's appearance consistent from reference images and supports four workflows at once — text-to-video, image-to-video, video-to-video and multi-reference — so it fits both generating from scratch and reworking existing footage.

Capabilities

  • Up to 4K video: crisp, detailed frames suitable for large screens, ads and commercial clips.
  • Native audio: the finished clip comes back with a soundtrack already synced to the on-screen action, so there is no need to source music or effects separately.
  • Character consistency: upload one or several reference images and your hero keeps the same face, outfit and features from shot to shot.
  • Four modes: text-to-video to build from scratch, image-to-video to animate a still, video-to-video to restyle a clip, and multi-reference for complex composite scenes.

How to try it

Open the Gemini Omni Video generation page, describe your scene — a detailed English prompt usually gives the most accurate result for video models — attach character reference images if needed, and start the job. The clip is ready in a few minutes and can be downloaded in original quality with sound.

When to use it

The model shines when you need a finished, sound-ready clip in a single step: ad inserts, product demos, short stories with a recurring character, or animating a static image. Video-to-video lets you reshoot an existing clip in a new style while keeping the original composition and motion, and multi-reference brings several characters or objects with a fixed look into one scene.

FAQ

What resolution is it, and does it have sound? The model generates video up to 4K and adds a native audio track straight from the model — no separate voice-over is required.

Can I keep the same character? Yes. Upload one or more reference images of your hero and Gemini Omni Video will hold their appearance across different scenes and angles.

How much does a clip cost? The cost is charged in credits and depends on duration and resolution; the exact amount is always shown before you start, so there are no surprise charges.

Try this model right now — no setup required.

Try Gemini Omni