OpenAI Unveils Sora, Its Text-to-Video AI Model

OpenAI has launched Sora, a text-to-video AI model available to ChatGPT subscribers, offering features like storyboarding, video remixing, and AI-powered scene blending, while emphasizing ethical use and creative potential.

Example Sora generation

OpenAI officially introduced Sora, its text-to-video AI model, as part of its “ship-mas” product release series. Sora, now live on Sora.com, is available for ChatGPT subscribers in the US and many other countries. Alongside Sora, OpenAI launched an updated model, Sora Turbo, which offers features such as video generation from text prompts, animating static images, and remixing existing videos.

For ChatGPT Plus subscribers, Sora allows the creation of up to 50 priority videos, using 1,000 credits, with resolutions up to 720p and five-second durations. The more robust ChatGPT Pro plan, priced at $200 per month, supports unlimited generations, higher resolution at 1080p, and videos lasting up to 20 seconds. Pro users also gain benefits such as watermark-free downloads and the ability to process up to five video generations simultaneously.

The Sora project has been in the works since its first tease in February. The launch gained further visibility through tech creator Marques Brownlee, also known as MKBHD, who previewed the model’s capabilities following hands-on testing.

During a livestream event, OpenAI highlighted key features of Sora. Its “storyboards” functionality enables the creation of videos based on a series of text prompts, while the “remix” tool allows users to refine Sora’s outputs with additional text. Another notable feature lets users “blend” two scenes seamlessly with AI. OpenAI also introduced an explore page, showcasing a community-generated feed of AI-created videos, making it accessible for browsing even to those without a subscription.

OpenAI emphasized ethical considerations around the use of Sora. Generated videos include visible watermarks and embedded metadata to indicate they’re AI-made. Before uploading any media, users must agree to terms prohibiting content involving minors, explicit or violent material, or copyrighted works. OpenAI’s moderation policies aim to prevent misuse while fostering creative expression. Addressing potential challenges, Sora product lead Rohan Sahai acknowledged that striking the right balance will require ongoing refinement and user feedback.

However, Sora’s rollout is not without limitations. CEO Sam Altman indicated that while the model is accessible in the US and other countries, availability in Europe and the UK may take additional time. The launch also comes amid controversy, as a group of artists previously involved in Sora’s alpha testing alleged that their contributions were used for unpaid R&D and marketing efforts.

Despite these hurdles, Sora’s debut marks a significant step forward in OpenAI’s efforts to expand creative tools powered by AI. Whether through exploring community content or leveraging advanced features like storyboarding and scene blending, Sora opens up new possibilities for video creation in the AI era.