VEO, Google’s AI capable of generating videos from text and images, is now available

Google is making significant strides to establish itself as a key player in the generative AI market, and its latest move is a major development. The tech company has just introduced a private preview of VEO, its video-generating AI, through Vertex AI, which is part of Google Cloud’s AI platform. In addition, Google has announced that Imagen 3, its most advanced image-generating AI, will be accessible to all Vertex AI customers starting next week.

VEO: Realistic, Coherent, and High-Resolution Videos

VEO, developed by Google’s DeepMind division, enables the creation of high-quality videos from text prompts or by using existing images as references. Clients of Google Cloud can use this AI to enhance their creative processes and make video production more efficient, achieving dynamic and professional results.

This AI can produce videos with a resolution of 1080p in a variety of cinematic and visual styles. Thanks to its sophisticated understanding of natural language and visual semantics, it can accurately follow user instructions. It also excels at maintaining coherence and consistency in videos, realistically recreating movements of humans, animals, and objects. While initially, Google stated that VEO could generate videos up to one minute long, it has not clarified whether the new private preview maintains this length or extends it.

PROMPT: The dog stands up, wagging his tail, happy and looking at the camera. pic.twitter.com/NZp5sTLGfW

— Google Cloud Tech (@GoogleCloudTech) December 3, 2024

Imagen 3: Precision and Detail in Image Generation

Starting next week, Google Cloud customers can begin using Imagen 3 from Vertex AI to create realistic and high-quality images. Google claims that this AI is their “highest quality text-to-image model, producing incredibly detailed and photorealistic images, with far fewer distracting visual artifacts than our previous models.”

Imagine: Generating high-definition images from a simple text prompt. Now you can with Imagen 3 on Vertex AI!

Check out some examples of Imagen 3’s image generation capabilities created from these prompts ↓
— Google Cloud Tech (@GoogleCloudTech) December 3, 2024

New Editing Tools

Additionally, Google has introduced a range of new editing and customization tools designed for commercial use. These editing tools are straightforward and allow for the refinement and customization of any image using written instructions. The “Inpaint” function is used to modify or add elements to an image, “Outpaint” lets you expand the image while maintaining context, and “Product background” helps change scenarios and backgrounds for product images.

On the other hand, “Imagen 3 Customization” offers greater control over the results. As Google explains: “Now you can incorporate your brand, style, logo, theme, or product features when generating new images. This opens up new creative possibilities, accelerating the development of marketing materials and advertising assets.”

Right now, these features are only available to a small, select group. However, if you’re interested, you can apply by filling out this form.

Google Overtakes OpenAI in Video Creation

Google introduced both VEO and Imagen 3 at its I/O 2024 event last May. The rush to make these technologies available to the market was not by chance, particularly when it comes to their video AI.

OpenAI, a significant competitor in this space, has yet to release Sora, its video-generating AI, to the public. Some companies, however, have been granted early access. A notable example is Toys ‘R’ Us, which has used Sora’s AI for their advertisement.

Photo: Google