OpenAI has officially launched Sora, its text-to-video AI model, as part of its 12-day “ship-mas” product release series. As previously reported by The Verge, Sora is now available on Sora.com for ChatGPT subscribers in the U.S. and “most other countries.” The release includes an enhanced version called Sora Turbo, which introduces features such as text-to-video generation, image animation, and video remixing.
OpenAI states that with a ChatGPT Plus subscription, users can create up to 50 priority videos (1,000 credits) at 720p resolution with a duration of up to 5 seconds. The recently launched $200 per month ChatGPT Pro subscription offers “unlimited generations,” up to 500 priority videos, enhanced 1080p resolution, and extended durations of up to 20 seconds. Pro subscribers also gain additional benefits, including watermark-free video downloads and the ability to generate up to five videos simultaneously.
OpenAI first hinted at its text-to-video AI model, Sora, back in February. Earlier today, Marques Brownlee, also known as MKBHD, confirmed its launch and shared a preview, highlighting his experiences testing Sora so far.
In the livestream, OpenAI showcased Sora’s new explore page, featuring a feed of AI-generated videos created by community members. The presentation highlighted a “storyboards” feature that enables video creation from a sequence of prompts, along with the capability to transform photos into videos. OpenAI also demonstrated a “remix” tool for refining Sora’s output using text prompts and a blending feature that seamlessly merges two scenes with AI.
OpenAI states that videos created with Sora will include visible watermarks and C2PA metadata to indicate they were AI-generated. Before uploading an image or video to Sora, users are required to agree that their content does not feature individuals under 18, explicit or violent material, or copyrighted content. OpenAI warns that “misuse of media uploads” may lead to account suspension or a ban.
“Sora product lead Rohan Sahai acknowledged during the livestream, ‘As OpenAI, we clearly have a big target on our back. Our goal is to prevent illegal use of Sora while also fostering creative expression. We understand this will be an ongoing challenge, and we may not get everything perfect from day one. We’re taking a slightly conservative approach to start, so if our moderation falls short, we encourage users to provide feedback.’”
Even without a ChatGPT subscription, you can browse the feed of AI-generated videos created by others using Sora. While the model launches today in the U.S. and many other countries, OpenAI CEO Sam Altman noted that its rollout in “most of Europe and the UK” may take some time.
Sora’s release follows a recent controversy where a group of artists, claiming to have participated in the company’s alpha testing program, leaked the product in protest. They accused OpenAI of utilizing them for “unpaid R&D and PR.”
