Google’s Veo 2 AI video model can now generate videos twice as long

Google DeepMind, Google’s AI research division, has unveiled a new AI model dubbed Veo 2 that rivals OpenAI’s Sora video creation tool. The Veo 2 text-to-video model succeeds the original Veo, which was announced back in May this year. Veo 2 has been improved with several features, including the ability to generate longer videos.

Veo 2 can also generate cinematic effects and output videos with SynthID watermark. Here’s everything you need to know about Google’s Veo 2 AI video model and how you can access it right away.

Everything you need to know about Veo 2

Google Veo 2 brings many updates over Veo. It can generate high-resolution videos across a wide range of topics. Videos generated by Veo 2 have more realism, and the tool appears to have a better understanding of physics, including human movement.

The Veo 2 AI model can also generate cinematic effects in videos. Users can specify if they want a wide-angle shot or the cinematic background blur in their videos. Google said that Veo 2 didn’t create any weird videos in its testing, which includes extra human fingers or distorted faces, meaning that there are fewer cases of hallucinations.

Unlike Veo, which was able to generate one-minute-long videos at 1080p resolution, Veo 2 can generate two-minute-plus clips at 4K (4096 x 2160 pixels) resolution. This is 4x the resolution and over 6x the duration of what Open’s Sora can do in terms of video generation.

Similar to the original Veo, the new Veo 2 AI model will output videos with an invisible SynthID watermark, ensuring that they are identifiable as AI-generated. This is great as it will  help avoid misinformation.

ALSO READ: OpenAI’s Sora announced for ChatGPT Pro, Plus subscribers

Veo 2 is now available to select users via the Google Labs VideoFX platform. It’s still not available in India, but we expect its availability here sometime later.

Other than this, Google also unveiled Imagen 3, which is its latest image generation model that produces more vibrant and better images.

Imagen 3 can generate images of different art styles, including photorealism, abstract, anime, and impressionism. The tool can be accessed in ImageFX and is currently available in over 100 countries.