Google Rolls Out Video AI Model for Gemini Users, Developers

Google has announced that its latest text-to-video AI model, Veo 2, is now available on Gemini and Google AI Studio.

To generate a video on Gemini, one must select the Veo 2 option from the model dropdown menu. This will generate an eight-second video clip at 720p resolution, which is available as an MP4 file with a 16:9 aspect ratio. 

As per the blog post, there is a monthly limit on the number of videos users can create, and they will be notified when they are close to it.

Google encourages users to provide all the required details for better control over the final video.

The video generation model is available only to Gemini Advanced users on the web and mobile. Subscribers can access this feature in all supported languages.

Developers can experiment with its capabilities in Google AI Studio and build things using a paid tier in the Gemini API. The blog post provides tips on crafting effective prompts for Veo 2, which encourages users to focus on clarity and details and include subject, action, setting, camera angle or movement, lighting, style, and mood for a promising output.

Not limited to Gemini and Google AI Studio, users can also use the model through Whisk. In this generative AI experiment, they can generate new images using text and image prompts and animate them into videos. To access this experimental tool, one must also be a Google One AI Premium subscriber.

The company also mentioned that all videos generated with Veo 2 are marked with SynthID, a digital watermark embedded in the video frames.