When OpenAI presented the 
Sora for the first time around 10 months ago, the world was amazed. Such plausible, artificially generated AI videos had never been seen before, nor were they expected so soon. And then it was time to wait, because OpenAI did not even give an approximate date as to when Sora would be available to everyone.
This meant that not only months passed, but also numerous competitors, who have shown ever better AI video models in recent months - which were always compared with the few Sora videos available to date.
OpenAI Sora before the release?
Now there are increasing indications that a release of Sora really is just around the corner. Whether this is already a version 2 is uncertain and also irrelevant, because OpenAI primarily wants to prove its reputation as the best state of the art (SOTA) AI video generator. To this end, there are still some open issues to be addressed, one of which is “character consistency”.
A leaked video from the C21Media keynote in London demonstrates that this problem may soon no longer be a problem:
As the video was filmed from a screen with a cell phone, not too much can be said about the quality, but OpenAI now seems to be able to keep characters stable over longer periods of time. It should also now be possible to create clips of up to 1 minute in length.
So will Sora be in the public spotlight in just a few days or hours? We think so - and would like to point out that OpenAI has only just presented its latest language model O1. Its use costs a self-confident 200 dollars per month. It is therefore not unlikely that Sora will also be priced in this range.
;)