OpenAI unveils Sora, its new text-to-video AI model

Microsoft-backed OpenAI, on Thursday, unveiled its new text-to-video model Sora, which can generate videos up to a minute long based on provided text prompts.  

Sora, currently accessible for red teaming purposes to identify AI system flaws, is also open for use by visual artists, designers, and filmmakers seeking feedback on the model, as stated by the company. 

Introducing Sora, our text-to-video model.

Sora can create videos of up to 60 seconds featuring highly detailed scenes, complex camera motion, and multiple characters with vibrant emotions. https://t.co/7j2JN27M3W

Prompt: “Beautiful, snowy… pic.twitter.com/ruTEWn87vf

— OpenAI (@OpenAI) February 15, 2024

“Sora is able to generate complex scenes with multiple characters, specific types of motion, and accurate details of the subject and background,” the statement said, adding that it can create multiple shots within a single video. 

In addition to its primary function of video creation from text prompts, Sora can also animate a static image, as highlighted in a blog post by the company. 

OpenAI, known for its popular chatbot ChatGPT and text-to-image generator DALL-E, is at the forefront of the generative AI revolution that emerged in 2022. The company emphasized in a blog post that Sora excels in accurately generating multiple characters and various motion types. 

OpenAI's broader goal, according to the post, is to train models that can comprehend and simulate the dynamic aspects of the physical world, aiding in problem-solving requiring real-world interaction. 

The introduction of Sora follows OpenAI's release of the ChatGPT chatbot in late 2022, which gained attention for its ability to compose emails, write code, and create poems. 

Meta, the social media giant, also enhanced its image generation model Emu last year, introducing two AI-based features for editing and generating videos from text prompts. 

While Sora is still a work in progress, OpenAI acknowledges that the model may struggle with spatial details in a prompt and encounter challenges in following specific camera trajectories. The company is actively working on tools to identify videos generated by Sora.