Gemini Advanced Subscribers Can Now Generate Video Clips With Veo 2

Sign in to your Android Police account

Summary
- Google is rolling out Veo 2 in Gemini Advanced, allowing subscribers to create 8-second videos from text prompts.
- Veo 2 prioritizes realistic physics and human anatomy to reduce visual abnormalities in generated videos.
- Veo 2 will also be available in Whisk, Google’s AI image generation tool, for creating short video clips from static images.
Last year, Google teased its latest AI video generation model, Veo 2. Early output looked impressive, with Google highlighting clips that, at a glance, were nearly indistinguishable from real footage. Now, Google is starting to roll out video generation in Gemini using the Veo 2 model to Gemini Advanced subscribers, allowing paid users to create short clips from simple text prompts.
Starting today, Gemini Advanced subscribers are able to generate eight-second-long, 720p clips by selecting Veo 2 from the model drop-down in the Gemini app and describing the video they’d like Gemini to create. Google has said that Veo 2 allows for highly specific video output based on your prompts, which can include directions about fine details, visual style, shot framing, and even simulated camera lens types. Generated videos come in a 16:9 aspect ratio and can be downloaded and shared as MP4 files.
Google’s said that Veo 2 can better represent realistic physics and human anatomy, which means that video output should exhibit fewer abnormalities like hands with extra fingers or objects behaving in unrealistic ways. Google says that each video frame generated by Veo 2 incorporates SynthID, an invisible watermark that identifies the content as AI-generated. SynthID won’t stop your average low-AI-literacy user from believing an AI-generated video is real, of course, but it’s something.
Veo 2 is also coming to Whisk, Google’s AI image generation tool that creates visuals based on three inputs: subject, scene, and style. Beginning today, Google One AI Premium subscribers can generate eight-second video clips based on images generated within Whisk, powered by the same Veo 2 model that will soon be available in the Gemini app.
Veo 2 is rolling out in Gemini Advanced starting today
It may not have all the capabilities Google initially talked up in its Veo 2 announcement — clips are limited to eight seconds in length, and resolution is fixed at 720p rather than the full 4K Veo 2 is capable of — but video generation in Gemini may be the most compelling reason to pay for Advanced today. Given how resource-intensive AI video generation is, I wouldn’t expect to see this feature make its way to the free Gemini experience any time soon.
What’s your reaction?
Love0
Sad0
Happy0
Sleepy0
Angry0
Dead0
Wink0
Leave a Reply
View Comments