Osegi
AI·1 min read

Google's Gemini Omni Now Generates Videos Up to 15 Seconds

Google's Gemini Omni model can now produce videos of up to 15 seconds, using image, audio and short motion references as input, the company said.

Illustration of a video timeline with image, audio and motion reference inputs feeding into an AI video generator.
Image: techinasia.com
SUMODialoediviewergiftdetect

Google's Gemini Omni can now generate videos up to 15 seconds long. According to the company, the model accepts image, audio and short motion references to guide what it produces.

What the update covers

The announcement centers on two points: the clip length, capped at 15 seconds, and the kinds of reference material the model can work from. Rather than relying only on a text description, Gemini Omni can take a still image, an audio track or a brief motion clip as a starting point for the output.

Reference-based input is a common request among people who make video with AI tools, because it gives them more control over how a scene looks, sounds and moves than a prompt alone.

Why it matters

A 15-second ceiling puts the feature in the range of short-form clips, such as social posts and brief promotional spots. Support for audio and motion references suggests Google is positioning Gemini Omni as a tool for guided creation, not just open-ended generation.

Google did not provide further details in the material available, so questions such as pricing, regional rollout, resolution and which Gemini plans include the feature remain open.

Apps worth a look

SUMO: the number puzzle

Tap the right tiles, add them up, hit the target. Easy to learn, hard to master. Free, for all ages.

Tags:googlegemini omnivideo generationgenerative aigemini

More from this day

See all →

More in: AI

See all →