OpenAI launches Sora, hyperrealistic AI video generator

OpenAI announced Public release of hyper-realistic AI video generation software sola Today—nearly 10 months after its first public display in February 2024.

In fact, OpenAI is actually releasing a model that’s significantly upgraded from the model that debuted at the time: the new Sora Turbo is due to SolaNet ChatGPT Plus and Pro paid subscribers ($20/month or $200/month) available in the US and most countries outside the EU and UK.

OpenAI co-founder and CEO Sam Altman introduced the news during a YouTube livestream as part of the company’s “12 Days of OpenAI” series of holiday-themed announcements scheduled for 1 p.m. ET / 10 a.m. PT .

Sora can generate a variety of videos from text input or still images, creating clips between 10 and 20 seconds in length, in resolutions ranging from 480p to 1080p, and in aspect ratios from landscape to square and vertical.

OpenAI has created a new and unique interface for the product, which includes a grid or list view that users can switch between to view their generations.

Users can also enter a mode called Storyboard, which allows them to generate multiple linked clips in the Timeline view. The model attempts to provide seamless transitions between clips—users can drag to make a clip more abrupt or to make it longer and smoother.

ChatGPT Plus users can generate up to 50 480p resolution videos per month.

For professionals and heavy users, the Pro plan offers higher resolution, longer duration, and low-speed unlimited generation.

OpenAI also announced plans to release customized pricing options for different user needs in early 2025.

News reported by MKBHD

Popular tech review YouTuber Marques Brownlee, better known as MKHBD, Break the news Sora was released about an hour early.

“The rumors are true – OpenAI’s artificial intelligence video generator SORA will be released to the public today…” Brownlee wrote in an article Post on social network.

Brownlee also shared an example of a video he made using the Text/Image/Video-to-Video Generator, to which he received early access as one of dozens of early creative partners OpenAI seeded ahead of the program’s official release.

Brownlee shared that while Sora can produce impressive and sometimes eerily realistic footage, such as that of newscasters or gadget commentators like himself, it also tends to produce hallucinations, random details and artificial intelligence generation. Obvious signs such as gibberish, meaningless text, unnatural physical phenomena, and even seemingly random additions or deletions of objects in the news chyron.

He also pointed out that OpenAI has fairly strict safeguards against generating portraits of real people, violence and explicit themes.

Nevertheless, in his complete YouTube commentshe finally concluded: “Human beings now have too many things to digest…[it] It’s the new baseline and it’s the worst ever.

Leaked on Hugging Face due to protests from early testers

This version comes after a leak Sora logs into AI code sharing community Hugging Face Beta testers launched such a protest about two weeks ago to protest OpenAI’s handling of the beta testing program. As the leaker wrote in their Hugging Face space:

“Hundreds of artists contribute pro bono labor to a $150B company through bug testing, feedback, and experimental work. While hundreds contribute for free, a select few will be selected through a competition to screen their Sora-created films— The compensation offered is minimal, which pales in comparison to the tremendous PR and marketing value OpenAI receives.

Sora is also in an increasingly competitive environment for generating realistic, live-action AI videos. track It continues to rapidly upgrade its AI video generation platform, rolling out new features, including the ability to re-record dialogue from existing footage and match characters’ faces, launched last week. forward artificial intelligence and Chinese competitors such as Klin, Heroinand recently, TencentIn the past few weeks alone, they have both launched impressive AI video generation tools.

So while OpenAI may have gained strong recognition with its success with ChatGPT and early attention-grabbing Sora footage that helped promote this new AI video generator to the masses, there are now a number of competing options, at least On the surface, it offers similar or better video quality. This makes Sola less of a guarantee of success.