The wait is finally over. Google has announced it’s rolling out AI-powered video generation in its Gemini app. Gemini Advanced subscribers can now get their hands on the company’s latest video model, Veo 2. First unveiled in December 2024, Veo 2 is designed to create highly realistic short video clips from text or image prompts.
Veo 2 in Gemini lets you create highly realistic short video clips
That means you can now create eight-second, 720p videos from text prompts, with results delivered as MP4 files in a 16:9 landscape format. To use the feature in Gemini, subscribers simply select Veo 2 from the model dropdown menu on the web or mobile apps, then describe the scene they want to generate. For example, you might write, “Generate a clip of a dog dancing in a bar.” Sure, that’s a wildly unrealistic scenario. But that’s exactly the kind of thing I’d be curious to see Veo 2 tackle.
Google recommends writing detailed prompts to get more accurate and creative results. Each video takes about 1–2 minutes to generate and can be downloaded or shared via a unique link.
Although Veo 2 is technically capable of generating longer and higher-resolution clips, Gemini Advanced users are currently limited to eight seconds at 720p. There’s also a monthly cap on video generation, but Google hasn’t revealed the exact limit. Users will be notified as they near their quota. On the safety side, every frame includes a SynthID digital watermark, clearly marking it as AI-generated content. Personally, I feel that’s a smart move in an age where synthetic media is everywhere.
Beyond Gemini, Veo 2 is also integrated into Google Labs’ Whisk platform. This lets Google One AI Premium users animate the images they create in Whisk using the same Veo 2 technology. A new “Animate” button allows users to describe how a static image should move, producing short, vivid video clips. The feature is rolling out globally today in all Gemini-supported languages.
