Google unveiled the VEO 2 artificial intelligence model to create video • DigiKala Mag

Google was very prolific last week and released numerous news about artificial intelligence, including news about “Gemini 2”, “Deep Research” and new features “Notebooklm”.

It is now time for the more creative part of this search giant. Deepmind has unveiled a new version of its VEO video model. VEO, first introduced at the Google I/O conference earlier this year, is a direct rival to “Openai” and is aimed at becoming one of the best artificial video manufacturers.

VEO 2, by improving visual realism and better understanding of physics, shows the movements more accurately. This is similar to the updates made by Pika Labs in the new Pika 2.0 model. According to Google, VEO 2 achieves very advanced results compared to other leading models, especially in the context of human faces.

«VEO 2What is the difference?

Google claims that VEO 2 can understand real -world physics. This is a ultimate goal for artificial intelligence video models and is an area that even the best, including “Sora” from Openai, is challenged. VEO 2 also has a good understanding of different types of cameras. This is the feature that the image models have been enjoying for a while and can be used effectively.

According to Google, you can: “suggest” 18mm lens “in your command and” VEO 2 “knows how to make the open angle view that this lens is known for.” You can also add the “low depth of field” to fade back. VEO 2 can produce clips for up to a minute with 4K resolution. This model is based on “Cinema Language”, and Google claims that you will see extra fingers and less unwanted objects in the videos.

VEO 2 has been added to Videoofx, but the service still has a waiting list. YouTube Shorts will also be added in the future to allow for the production of artificial intelligence content on this video platform.

Source: 9to5Google