OpenAI Introduces New Text-To-Video AI! Here's What Users Need To Know About Sora

OpenAI introduced its newest text-to-video AI model called "Sora." On Thursday, Feb. 15, the AI developer unveiled a demo of Sora to show users how capable this new artificial intelligence tool really is.

OpenAI CEO Sam Altman boasted the new Sora AI model on X, saying that the arrival of it is truly a "remarkable moment." As of writing, Altman said that they are starting red-teaming and offering access to a limited number of creators.

This means that not everyone can access Sora. If you're among the excited AI fans, here's everything you need to know about this new text-to-video AI tool.

Sora: OpenAI Introduces New Text-To-Video AI!

(Photo : STEFANI REYNOLDS/AFP via Getty Images)
This photo illustration shows the ChatGPT logo at an office in Washington, DC, on March 15, 2023. The company behind the ChatGPT app that churns out essays, poems or computing code on command released on March 14, 2023.

Sam Altman confirmed that Sora's public expansion will soon come. But, he didn't provide an exact date. The arrival of Sora is a big deal for the AI developer.

This is because it allows OpenAI to make its first step in entering the intensely competitive space of text-to-video AI. Right now, numerous companies are already offering AI video generation models.

These include Pika, Runway, Stability AI, and other startups. But, unlike the models that OpenAI's rivals offer, Sora promises more advanced features.

One of these is run time. For example, the AI video generation models of Pika and Runway can only provide 4-second clips at a time. But, OpenAI's Sora can provide longer videos; its limit right now is 60 seconds.

Aside from run-time, Sora's videos stand out because of their resolution sharpness, human anatomical accuracy, motion smoothness, and physical world accuracy, as reported by Venture Beat.

Thanks to its advanced capabilities, Sora not only offers fictional videos-like wizards casting spells, dinosaurs fighting, etc.-but it can also provide insanely accurate real-life situations, such as couples enjoying the cityscape view, driving an off-road vehicle, and more!

Sora Still Needs More Work

(Photo : KIRILL KUDRYAVTSEV/AFP via Getty Images)
A photo taken on November 23, 2023 shows the logo of the ChatGPT application developed by US artificial intelligence research organization OpenAI on a laptop screen (R) and the letters AI on a smartphone screen in Frankfurt am Main, western Germany.

Although Sora is more efficient than other text-to-video AI models, OpenAI admitted that it still has some weaknesses.

The AI developer shared that Sora can't accurately simulate the physics of a complex scene. Another issue with Sora right now is it may not understand specific instances of cause and effect.

Aside from these, another problem with Sora is it may confuse the spatuial details of a prompt, such as mixing up right and left. But, despite being a new AI model, there's no denying that Sora is a promising innovation.

It even attracted the attention of Mr. Beast, who jokingly said to Altman that he's hoping Sora won't make him homeless.

"Sora serves as a foundation for models that can understand and simulate the real world, a capability we believe will be an important milestone for achieving AGI," said OpenAI via its official website.