OpenAI Releases Text-To-Video Generation Model ‘Sora’

On Monday, artificial intelligence giant OpenAI officially launched its highly-anticipated text-to-video generative AI model Sora. It was released as part of the company’s 12-day “Ship-mas” series and will be available starting today for ChatGPT subscribers in the U.S. on Sora.com. 

OpenAI Launches Text-To-Video Model ‘Sora’ To U.S. Customers 

Sora was first teased in February 2024, then accessible only to media professionals and OpenAI’s partners. The text-to-video model took the tech world by surprise with its ability to render and produce high-quality videos. However, instead of releasing it right away, OpenAI decided to continue building on Sora and perfect its capabilities. 

The company has also announced an advanced version of Sora, called the Sora Turbo. Users with a ChatGPT Plus subscription can generate up to 50 videos per month at resolutions of up to 720p with a duration of 5 seconds. The $200 ChatGPT Pro subscription, which launched last week, allows users to generate up to 500 videos per month at 1080p and durations of 20 seconds each.

While users can download the videos prompted under both subscriptions, those signed up for the premium one can download the clips without a watermark and perform up to five simultaneous generations. 

During the launch, OpenAI showed off Sora’s explore page, containing a feed of AI videos curated by other community members. One main feature highlighted by the company was “Storyboards”, which lets Sora users generate videos based on a sequence of prompts, and the ability to turn pictures into videos. Then there was the “Remix” tool that lets users change Sora-produced content with a text prompt and a “Blend” tool that merges two scenes. 

OpenAI Instills Robust Privacy and Copyright Rules For Sora

OpenAI specified that all videos generated with Sora will contain visible watermarks and C2PA metadata to indicate that they were produced by AI. Before uploading an image or video to Sora, users will be asked to check off a prompt box agreeing that the content is not explicit or violent, contains people under 18, or copyrighted material. Any misuse of the service could result in the account being banned or suspended. 

CEO and co-founder Sam Altman described Sora as the “GPT-1” of video production, emphasizing how it is still in its infancy. He added that despite this, Sora’s feed is “compelling” due to the kind of videos uploaded by subscribers. The feed is open to everyone, and the clips can be played without needing a ChatGPT account. 

Sora’s launch comes during a cumulative period for OpenAI. The text-to-video model was leaked last week after a group of artists, who claimed to be part of the testing program, accused the company of “unpaid R&D and PR”. 

Elon Musk vs OpenAI

The AI giant, which started as a non-profit organization, has transitioned into a profit-generating outfit and secured funding from investors to double down on its product research and development. 

The company is embroiled in a lawsuit with Elon Musk, an angel investor in the company, who is seeking to halt its plans to become a for-profit entity. The Tesla and SpaceX CEO has also filed an antitrust lawsuit against OpenAI and its parent company Microsoft, accusing them of anti-competitive practices and conflicts of interest in the AI industry. 

Musk’s lawyers accused both firms of engaging in market manipulation tactics by deterring investments aimed at rival AI ventures, including his own xAI. The lawsuit alleges that OpenAI coerced investors into not supporting its competitors that have recently secured funding.

Also Read: XYO(XYO) price prediction (Where to Buy , How To Buy) (Is it a good investment for the short term?)