OpenAI has released Sora, its text-to-video generation model. This tool allows users to breathe life into their imagination by creating realistic videos from text prompts.
Users will be able to generate videos up to 1080p resolution for up to 20 seconds. The platform supports various aspect ratios, including widescreen, vertical, and square. Additionally, users will also be able to add and integrate their own assets into the videos.
Sora is a standalone product available on Sora.com. It is currently offered to ChatGPT users in the US who have a Plus or Pro subscription as a free add-on.
ChatGPT Plus account holders will be able to generate up to 50 videos at a resolution of 480p each month, or fewer videos at a resolution of 720p. In contrast, Pro plan users have access to up to ten times more usage, with much higher resolutions and longer video durations. More plans will be added in the future for different user needs.
To address potential concerns of misuse, OpenAI has implemented strict moderation standards for uploaded content featuring people. Additionally, the platform will continue to block the generation of content with nudity.
OpenAI has also worked with “red-teamers,” security experts specializing in disinformation, illegal content, and online safety, to conduct testing of the model to identify potential risks associated with its use.
Sora is a significant development in AI-generated content, though its potential impact remains to be seen.
