Google Veo, a new artificial intelligence model for generating videos, has become available to companies, who can now connect it to their content creation applications. OpenAI presented a similar Sora three months earlier, but it is still not possible to work with it, while Veo is already available in preliminary access on the Google Vertex AI platform.
Image source: Google
Veo supports the generation of “high quality” 1080p video in a variety of visual and cinematic styles for queries in text or image format. When the model was announced in May, these videos were slightly longer than a minute in length, but now Google does not specify any restrictions. The quality of the samples is truly impressive: you need to look very closely at the video to see signs of AI.
In addition to Veo, all Google Cloud clients “starting next week” will be able to connect the Imagen 3 image generator through Vertex AI. Some will have access to additional functions: editing photos using text queries or the ability to add your own corporate style to the generated image. Veo and Imagen 3, Google says, have built-in protections that will prevent the creation of harmful content or copyright infringement. Materials created by both models receive the SynthID label, indicating the involvement of generative AI.
With Google already bringing its products to market, OpenAI is lagging behind competitors and has less time left to fulfill its promise to release Sora by the end of 2024.