Attempting to stay ahead of OpenAI in the generative video race, Google announced Veo 2, which it says can output 4K clips of two-minutes-plus at 4096 x 2160 pixels. Competitor Sora can generate video of up to 20 seconds at 1080p. However, TechCrunch says Veo 2’s supremacy is “theoretical” since it is currently available only through Google Labs’ experimental VideoFX platform, which is limited to videos of up to 8-seconds at 720p. VideoFX is also waitlisted, but Google says it will expand access this week (with no comment on expanding the cap).

Google DeepMind VP Eli Collins tells TechCrunch that Google is making Veo 2 available through its Vertex AI developer platform “as the model becomes ready for use at scale.”

The competition is moving at light speed. Google released Veo in private preview on December 3, trying to get a jump on OpenAI as work leaked of Sora’s imminent public availability, which came on December 9.

As with its predecessor, Veo 2 can generate videos from either a text or image prompt — including from pictures generated by Google’s Imagen image generator, simultaneously upgraded to Imagen 3.

Imagen 3 is available in ImageFX and Google Labs’ new AI experiment, Whisk, described by SiliconANGLE as “a remix tool” that “allows users to combine multiple existing images into a new one. Whisk can, for example, apply one the style of one image to another’s background.”

As to what it has improved in the two weeks since Veo debuted in preview, Google DeepMind says the model can better understand physics and camera controls while also producing “clearer” footage.

“By clearer, DeepMind means textures and images in clips are sharper — especially in scenes with a lot of movement,” TechCrunch writes, explaining that the camera control upgrades “enable Veo 2 to position the virtual ‘camera’ in the videos it generates more precisely and to move that camera to capture objects and people from different angles.”

DeepMind also says Veo 2 has more realistic fluid dynamics (liquid pouring), and better handles lighting effects, like shadows and reflections. “That includes different lenses and cinematic effects, DeepMind says, as well as ‘nuanced’ human expression,” TechCrunch reports, viewing the improvements appreciatively, while noting “Veo 2 can’t quite clear the uncanny valley.”

“While Veo 2 is available only to select users, the original Veo remains available on Vertex AI,” VentureBeat reports, noting that “videos created with Veo 2 will contain Google’s metadata watermark SynthID to identify these as AI-generated.