Gemini Omni Video — AI Video Generator

About Gemini Omni Video

Features

• Accepts combined text, image, audio and video inputs in one prompt, including sketches and style or motion references.
• Generates clips of up to 10 seconds at launch, a cap Google describes as a deployment decision rather than a model limit.
• Multi-turn conversational editing: swap characters or objects, change camera angles, transform environments, or apply motion and style from reference material across successive turns.
• Edits real uploaded footage as well as generated clips, changing the action inside a video you shot.
• Holds character consistency and scene physics across sequential edits.
• Supports 16:9 cinema and 9:16 vertical framing, with audio synchronized to on-screen action.
• Every output carries an imperceptible SynthID watermark, verifiable through the Gemini app, Chrome and Google Search.

What people use it for

• Turning a rough brief of mixed media, such as photos plus a voice note, into a finished short clip.
• Editing real footage by prompt: adding characters, changing what happens, or restyling a moment without a timeline editor.
• Short-form content for YouTube Shorts, where Omni Flash launched cost-free in the Shorts and YouTube Create integrations.
• Iterative story development, refining one shot over many conversational turns instead of regenerating from scratch.
• Scenes that depend on plausible physics or factual detail, where Omni's world-knowledge grounding is the differentiator.

Reported strengths

• Input flexibility goes well past text-to-video and image-to-video; TechCrunch called it a concrete step toward Google's original natively multimodal Gemini vision.
• The conversational edit loop keeps characters and physics stable across changes, which one-shot video models struggle to do.
• Knowledge-grounded generation draws on Gemini's understanding of history, science and cultural context.
• Broad day-one distribution: Gemini app and Flow for AI Plus, Pro and Ultra subscribers globally, plus free YouTube Shorts access.

Reported limitations

• Clips are capped at 10 seconds in the launch rollout of Omni Flash.
• Editing prompts need to be very specific; TechCrunch reported that loosely worded edits can alter details you meant to keep.
• Audio-to-video input and image or audio outputs are staged for later, and speech synthesis for avatars is still under responsible-deployment review with identity verification required.
• API access trailed the consumer launch by weeks, and the higher-fidelity Omni Pro tier was unreleased at announcement.

Sources

1. Introducing Gemini Omni (Google)
2. Gemini Omni - Google DeepMind (Google DeepMind)
3. Google's Gemini Omni turns images, audio, and text into video - and that's just the start (TechCrunch)