Google unveils Veo 2 and Imagen 3, two powerful video and image generators, and Whisk, a content remixer

A few days after launching Gemini 2.0, Google has updated its repertoire of AI content creation tools with Veo 2, a powerful video generator with the potential to surpass Sora; Image 3, a static image generation model with abundant improvements; and Whisk, an interesting tool capable of remixing existing images using AI to create new compositions.

I See 2: Longer, Sharper Videos with Advanced Physics Effects

Starting with Veo 2, the new video generator developed by Google Labs surprises with its enormous evolution. Even though Veo was released less than a year ago, its revision is capable of delivering four times the resolution of Sora and six times the length of its clips, although in practice (and for now) they are limited to 720p and less than ten seconds in VideoFX.

In addition to providing higher resolution and being able to create longer short films, Veo 2 also incorporates very important improvements in composition, the definition of scenes with a lot of movement and in physics effects, including properties of liquids, reflections and objects in collision, although logically the coherence problems inherent to any AI generation model continue to persist.

Google has not indicated when or if Veo 2 will be available in Spain (VideoFX is only offered in the US, Australia, New Zealand and Kenya), but indicates that its plans include integrating it into YouTube Shorts. Meanwhile, the company has already published a waiting list.

Image 3: greater refinement and hands with the appropriate number of fingers (almost always)

As for Image 3, the third revision of the Google Labs static image generator offers notable optimizations in terms of style, composition and detail. Illustrations created using Image 3 are brighter and more vivid, moving closer to photorealism step by step.

Image 3, in any case, seeks to create more credible images, which are not necessarily close to reality. Google Labs has also refined creation in styles like anime and impressionism, all while reducing the appearance of artifacts, from smudges to six-fingered hands.

Image 3 integrates into ImageFX and is available globally.

Whisk: AI illustration generator from other images

Finally, Whisk is a new tool that allows you to generate images from existing images, without the need to enter instructions through text. Thinking about those users who do not even know how to use a prompt, Whisk allows you to select a subject from an image or description, a selected scenario using those same possibilities and a style to create new compositions.

Whisk uses the Image 3 image generator as a base and is for now an experimental tool only accessible through a waiting list that does not include Spain.