Open AI has unveiled a new version of its text-to-video generation model called Sora Turbo. Nearly 10 months after the company previewed the Sora model, it announced that the latest version would be available to Chat GPT Plus and Pro users via a designed website. The model can generate videos up to 20 seconds long at resolutions of 1080p from a text or image prompt.
During its Monday live-streamed presentation, Open AI said that Sora Turbo will be available in the US markets and other parts of the world but not yet in Europe. Existing Plus users who tried to access the tool were presented with a message that “sign-ups are temporarily unavailable due to heavy traffic.”
“We developed a new version of Sora—Sora Turbo— that is significantly faster than the model we previewed in February. We’re releasing it today as a standalone product at Sora.com to ChatGPT Plus and Pro users,” the company said in a statement.
The company added that the new Sora’s availability will be limited. While those who pay $20 monthly for a ChatGPT Plus subscription will have access to generate only 50 videos monthly, users who pay $200 for a ChatGPT Pro subscription will have unlimited access to “slower” generations and 500 videos created at a faster speed.
According to the OpenAI Chief Executive, Officer Sam Altman credited the video technology’s early stage but stressed that “it’s going to get a lot, a lot better.” He added that Sora won’t be available in most parts of Europe which have tougher tech regulations.
Another feature of the new tool is its ability to produce videos from still images, videos, and text prompts. The company explained that a user could upload a video clip and the software will spit out footage to lengthen it.
In addition, Sora Turbo also includes an option called Storyboard. This feature allows people to write out what they want to happen at different points in a video in order to better direct a sequence of clips. There’s also an option to explore a feed of videos created by other users.
OpenAI has been recently working on keeping pace with a growing number of tech companies that offer tools to produce short clips quickly and more cheaply than creating video from scratch. Though tech giants like Google and Meta have also announced similar video tools none have yet met the huge expectations set for AI since the launch of ChatGPT two years ago.
Sora – Concerned challenges
Following its launch in February, Sora had constraints after it made a series of mistakes. A report by Bloomberg News found Sora struggled with stray body parts and a mixed understanding of physics. It also required more time to generate each video than users might be accustomed to with an image generator or chatbot.
However, Open AI acknowledged technical limitations in the current release. “This early version of Sora will make mistakes, it’s not perfect,” said one developer during the live stream launch.
“The version of Sora we are deploying has many limitations. It often generates unrealistic physics and struggles with complex actions over long durations. Although Sora Turbo is much faster than the February preview, we’re still working to make the technology affordable for everyone,” the company added during its Monday presentation.
In its efforts to consider deepfake attacks, the platform implemented C2PA metadata in all generated videos for identification and origin verification. OpenAI also developed an internal search tool to verify Sora-generated content.
In addition to the release, OpenAI also published Sora’s System Card for the first time. The feature includes technical details about how the model works and safety measures the company undertook before this release. Sora also makes use of a recaptioning technique to generate highly descriptive captions for the visual training data. This allows the model to “follow the user’s text instructions in the generated video more faithfully,” OpenAI writes.
Sora Turbo’s announcement came as part of what OpenAI has said will be a 12-day of livestreamed product events. The company introduced an advanced-priced tier for its chatbot ChatGPT featuring unlimited access to its most powerful artificial intelligence designs. It said that the ChatGPT pro is set to cost $200 (N150,000) a month.
Open AI highlighted that the new advanced model has been upgraded to perform human-like reasoning tasks and offers access to an expanded version of its current reasoning model — o1. In addition to the latest Open AI software development, Open AI also offers a ChatGPT Plus subscription for $20 (N15,000) a month, a tenth of the new option’s cost.
The tech giant previously released a preview of the new model in September where the company said it has been enhanced to solve competitive math and code problems in a faster and better way.
Read More: Open AI launches ChatGPT pro that performs human-like reasoning, to cost $200 a month.
Technext Newsletter
Get the best of Africa’s daily tech to your inbox – first thing every morning.
Join the community now!
