Google unveils Veo 2 video-generation tool to rival OpenAI’s Sora: All the details

Veo 2

Google has introduced Veo 2, a new video-generation tool that can generate up to 4K resolution videos. This new model promises to create incredibly realistic videos that are almost indistinguishable from real-life footage. It comes just days after OpenAI made its video generation model, Sora AI, available to the public.

Google’s Veo 2 video-generation tool: Key details

The company has shared a series of short video clips on its social media handles and via a new blog post created using Veo 2, which shows that this platform can generate hyper-realistic videos of animals and food. It also has some key improvements over its predecessor, which was released in May this year. The new model now has a perfect understanding of human movement and facial expressions, which makes the output footage more realistic.

Moreover, unlike Sora’s capped limit of up to 20-second clips, another area where Veo 2 has a theoretical advantage is that it can generate a clip with a duration of over 2 minutes, which is more than 6x the duration of Sora. Google also claims that common hallucination effects, where most AI models tend to generate extra fingers or unexpected objects to ensure the resulting footage looks more realistic, have been rectified by Veo 2 when compared to the videos generated by other AI models.

Currently, Veo 2's access is limited to Google Labs' VideoFX users only, and new users can sign up for the Google Labs waitlist to get access to Veo 2. There are also plans to integrate Veo 2 with YouTube Shorts and other Google products in 2025. Alongside Veo2, Google has also introduced Imagen 3, which is a state-of-the-art image-generation tool that can create photorealistic images.