Beginning today, Google has opened access to Veo 2, its text-to-video AI model, for subscribers of Gemini Advanced. The company claims that Veo 2 can create high-resolution video clips with a quality akin to “cinematic realism.” Users can now choose Veo 2 from the Gemini model options on both web and mobile platforms, allowing them to enter a prompt for generating an eight-second video at a resolution of 720p.
Subscribers will encounter restrictions on the number of videos they can generate monthly, with notifications alerts provided when they near their limit. The output format for videos is MP4, and mobile users can directly upload their creations to platforms like TikTok and YouTube via a “share” button.
According to Google, Veo 2 benefits from improved capabilities that enhance its comprehension of real-world physics and human movements. This leads to more fluid character animations, lifelike settings, and enhanced visual details across a variety of subjects and artistic styles. Each generated video from Veo 2 includes SynthID digital watermarks, signaling their AI-generated nature.
Besides Veo 2, Google has also introduced Whisk Animate, a new feature that enables users to turn still images into eight-second videos using Veo 2. This tool is accessible to Google One AI Premium subscribers and expands on the capabilities of the existing Whisk tool, designed for AI-generated image mashups. Whisk Animate is available worldwide via Google Labs.
Initially, Veo 2 was launched in early access format back in December, with a cost of 50 cents per second for video processing on Google’s Vertex AI platform. YouTube has incorporated Veo 2 into its experimental Dream Screen feature, allowing users to create AI-generated video content for Shorts.
