Back in mid-February, OpenAI unveiled Sora, its generative AI model that creates videos from text prompts or even a still image and can also edit videos or extend them by filling in or adding frames. Amazing as the demos were, there was no solid launch date for the general public - all we were told was that Sora was being tested by experts who were carrying out tests and were assessing the model for potential risks ahead of launch.

Today thanks to an interview by OpenAI's Chief Technology Officer Mira Murati we know that it's only a matter of months - Sora will be released later this year. It could in fact be ready "in a few months".

Of course, that's far from a hard date, but it's good to know that this isn't several years away from an actual launch. When Sora arrives, it will also let you edit the videos it creates. Initially the videos will be devoid of any sound, but incorporating audio into them is on the roadmap for a later release.

According to Murati, Sora is much more expensive to run compared to other AI models, but even so OpenAI wants to make it available at "similar costs" to DALL-E, the image generating AI tool. In terms of safety and security, the videos generated with Sora will be watermarked and the model "likely" won't be allowed to generate videos of public figures.

Source