Google launches Veo 2: AI generates more impressive videos than Sora
After generative AI for images or text, it's time for models that shape videos. OpenAI revealed Sora and Google Deepmind enters the race with Veo 2 which has solid arguments.
Veo 2 has solid arguments against Sora
Today, we're announcing Veo 2: our state-of-the-art video generation model which produces realistic, high-quality clips from text or image prompts. 🎥
We're also releasing an improved version of our text-to-image model, Imagen 3 – available to use in ImageFX through… pic.twitter.com/h6ejHaMUM4
— Google DeepMind (@GoogleDeepMind) December 16, 2024
The major innovation of Veo 2? His understanding of cinematic language. The AI accurately interprets technical terms like 18mm for a wide shot or visual effects like shallow depth of field to blur the background.
Google Deepmind is proceeding with a cautious deployment despite these impressive advances. Access to Veo 2 is via the VideoFX platform which has set up a waiting list. The features are also limited with 720p and 8 seconds, maximum.
This is a different strategy from Sora. OpenAI allows ChatGPT Pro subscribers to generate 1080p videos up to 20 seconds.
But that's not all for the creative part. Google strengthens image generation with Imagen 3. The update is accessible via ImageFX with a Google account and allows renderings with precise artistic styles and fidelity of instructions.
