OpenAI releases the video generator Sora AI to the general public

The long-awaited Sora video text generator, powered by artificial intelligence from OpenAI, is set to be available to all users soon.

YouTube user Marques Brownlee has shared this news in a video on his channel. According to Brownlee, the Sora generator is now available on the official Sora website.

Sora can create videos up to one minute, maintaining visual quality and following user instructions. On the Sora website, it said "We're teaching AI to understand and simulate the physical world in motion, with the goal of training models that help people solve problems that require real-world interaction."

Sora is able to create complex scenes with multiple characters, a variety of movements, and detailed objects and backgrounds. The model not only considers user requests, but it also understands how these elements exist in the real world.

The current model still needs to be improved. For instance, there may be challenges in simulating the physical properties of a complex environment and understanding specific instances of causality. The model might also confuse spatial details provided in the input, or struggle to accurately describe events that occur over time, such as certain camera movements.

According to Brownlee, Sora can generate videos with a resolution of up to 1080 pixels, but the higher the resolution, the longer the video takes to create. For example, 1080-pixel videos take 8 times as long to load as 480-pixel ones, which are the fastest option. 720-pixel videos, meanwhile, take 4 times as long.

The full Sora release is expected by the end of the year.