Google announced Veo, a generative AI model that can create high-quality videos based on users' text input, at the Google I/O 2024 developer conference.
Google claims that Veo models can understand natural language and visual semantics in advanced ways to create any video that comes to mind.
According to Google, Veo can create videos longer than a minute at 1080p resolution, and is also capable of understanding cinematic and visual techniques such as the concept of time-lapse.
Google collaborated with director Donald Glover and Gilga Studios to demonstrate the Veo model's ability to simulate real-world physics, as shown in a promotional video the company posted on YouTube.
Veo templates will be available in Google's VideoFX tool for select creators starting today and will also be added to YouTube Shorts and the company's other products.
Google has also released Imagen 3, an advanced text image model, which the company says is a “high-quality” text image model, offering stunning detail, lifelike and realistic images, and some flawless images.
Google added that Imagen 3 models now better process text and understand details in long data.
On the other hand, the technology community is eager to try out Google's new model and compare its performance with competing OpenAI models, including the Sora text-to-video model and the DALL-E-3 text-to-video model. . Convert text to image.