Gemini AI can generate 720p videos up to 8 seconds long

Google has begun a gradual rollout of video generation functionality in Gemini Advanced. Now, subscribers can create short videos using the Veo 2 model by simply describing the scene in text format.

The new model allows you to generate videos up to 8 seconds long in HD 720p format and an aspect ratio of 16:9. The main emphasis is on cinematic, detailed and realistic animation – the model better recognizes the physics of movement and conveys the behavior of characters and objects in the frame. Thanks to this, scenes look more believable even with complex compositions.

How it works

To use Veo 2, the user needs to select it from the list of models in the Gemini interface. Next, enter a text query describing the desired video scenario. The more precise the description, the more manageable the result will be. It can be as simple as a visual idea or a mini-story.

There is a limit on the number of videos per user, which varies — Google does not disclose the exact value. However, the system will warn you in advance if you are approaching the limit.

The generated video can be saved and shared – for example, uploaded to TikTok or YouTube Shorts directly from the interface.

Each video is marked with a digital SynthID watermark embedded in each frame. It indicates that the video was created using AI. There are also restrictions on creating content that violates the platform’s rules.

The feature is already becoming available to Gemini Advanced users around the world – both in the web version and in the mobile application, but full implementation will take a few more weeks. Support covers all languages ​​available in Gemini.