OpenAI has officially launched its highly anticipated text-to-video AI model, Sora, as part of its ongoing 12-day “Ship-mas” product release series.
The new model, available today on Sora.com for ChatGPT subscribers in the US and most other countries, is designed to generate videos directly from text prompts. Along with the launch, OpenAI introduced a new version of the model, Sora Turbo, which includes several enhanced features such as animating images, remixing videos, and creating videos from text.
To access Sora’s advanced features, users need a ChatGPT subscription. With a ChatGPT Plus subscription, which costs $20 per month, subscribers can generate up to 50 priority videos (1,000 credits) at a resolution of 720p with video durations of up to 5 seconds.
For $200 per month, the ChatGPT Pro subscription offers unlimited generations, 500 priority videos, higher resolution (up to 1080p), longer video durations (up to 20 seconds), and additional perks like watermark-free downloads and the ability to perform up to five generations simultaneously.
OpenAI demonstrated Sora’s capabilities during a livestream event earlier today. Key features include a new explore page, where users can browse AI-generated videos created by the community, and a storyboard tool that allows users to generate videos based on a sequence of text prompts.
Another standout feature is the ability to turn still images into animated videos. OpenAI also unveiled a remix tool that lets users tweak and refine videos generated by Sora through text-based instructions. Additionally, a blend function allows users to seamlessly combine two scenes together using AI.
Despite its potential for creativity, OpenAI has put measures in place to ensure that content uploaded to the platform adheres to ethical standards.
Users are required to agree that the media they upload does not contain explicit, violent, or copyrighted material, nor does it feature individuals under the age of 18. Sora-generated videos will also include visible watermarks and C2PA metadata to indicate that the content was created by AI.
Sora’s product lead, Rohan Sahai, acknowledged the potential challenges in moderating content while balancing creative freedom. During the livestream, he noted, “We obviously have a big target on our back as OpenAI. We want to prevent illegal activity, but we also want to foster creative expression. We know this is an ongoing challenge, and we might not get it perfect on day one.”
Sora’s moderation system aims to prevent the misuse of the platform, but OpenAI has urged users to provide feedback if the system does not correctly filter content. Sahai emphasized that OpenAI is starting cautiously, with an emphasis on maintaining safety and compliance.
While Sora is now available in the US and many other countries, OpenAI CEO Sam Altman stated that the model’s availability in most of Europe and the UK may take longer. For those without a ChatGPT subscription, Sora’s explore page will remain accessible for browsing AI-generated videos created by other users.
The release of Sora comes on the heels of controversy. Last week, a group of artists who had allegedly participated in OpenAI’s alpha testing program leaked details about the product, accusing the company of using them for “unpaid R&D and PR.” The artists' protest centers around their claims that OpenAI has not compensated them for their work and contributions to the model's development.
