OpenAI rolls out text-to-video model Sora for ChatGPT Plus & Pro users

OpenAI has launched its advanced artificial intelligence video generation model, Sora Turbo, to ChatGPT Plus and Pro subscribers, broadening its capabilities in multimodal AI technologies. The move marks a step in OpenAI’s efforts to rival similar offerings from tech companies such as Meta, Google, and Stability AI.

Initially introduced in February during a research preview, Sora Turbo is now accessible at no additional cost to premium ChatGPT users. The model enables the creation of videos up to 1080p resolution, with a duration of up to 20 seconds, across various aspect ratios, including widescreen, vertical, and square.

Before its official release, the Sora model was leaked during its initial research preview earlier this year. Some users gained unauthorised access to the platform through technical vulnerabilities, leading to the circulation of test outputs online. The company addressed these leaks by tightening access controls and emphasising the need for responsible use during its research phase. The company has since implemented additional safeguards, including watermarks and metadata, to ensure transparency and security in its current deployment.

In a statement, OpenAI noted its plans for "tailored pricing for different types of users,” which it aims to implement in early 2025. The company has also rolled out a new interface for Sora, featuring tools such as frame-specific inputs via a storyboard and options to remix, blend, or generate entirely new content from text.

Key Features of Sora

• 1080p Resolution: Create videos in 1080p, up to 20 seconds long, with various aspect ratios.
• Remix: Modify video elements by replacing, removing, or re-imagining them.
• Re-cut: Isolate and extend specific frames to complete scenes.
• Storyboard: Organise and edit videos using a personal timeline.
• Loop: Trim and create seamless looping videos.
• Blend: Merge two videos into one cohesive clip.
• Style Presets: Apply and share creative video styles.

Safety and ethical measures

The company has implemented stringent safeguards to mitigate potential misuse of Sora. This includes blocking the creation of child sexual abuse materials and sexual deepfakes. At launch, uploads featuring real people are restricted, with plans to expand these capabilities as the platform refines its deepfake protections.

“All Sora-generated videos will include C2PA metadata to identify them as AI-generated,” OpenAI said, adding that visible watermarks are enabled by default. The company has also developed an internal search tool to verify content authenticity.

Sora Turbo’s availability currently excludes the European Union, Switzerland, and the United Kingdom, though it is accessible in other regions where ChatGPT operates.

Early adoption and future prospects

The company described Sora Turbo as “significantly faster” than the initial version unveiled earlier this year. While the technology remains limited in handling complex actions and long-duration videos, the company said it aims to refine the model and ensure its affordability for a broader audience.

The decision to release Sora Turbo now is part of the company's strategy to encourage societal exploration of the technology and establish norms and safeguards. The model is positioned as a tool for creative storytelling, enabling users to push the boundaries of video generation.

“This early version of Sora will empower users to explore new forms of creativity and expand the possibilities of video storytelling,” OpenAI said.

Users can access Sora at sora.com, with Pro subscribers benefiting from higher resolutions, longer durations, and greater usage limits.