OpenAI launched Sora, its text-to-video generation model, to ChatGPT Plus and Pro subscribers on 9 December 2024 [source]. The release follows months of limited testing with creative professionals and marks the first public availability of the system.

Sora generates video clips up to 20 seconds at 1080p resolution from text prompts. ChatGPT Plus users can generate up to 50 videos per month at 480p resolution with watermarks, while Pro subscribers receive higher limits and access to 1080p output without visible branding. The model supports aspect ratios including widescreen, vertical, and square formats.

OpenAI stated the system includes C2PA metadata for provenance tracking and employs content moderation filters to block requests depicting minors, public figures, copyrighted characters, and explicit material. Videos are screened before generation and after completion. The company acknowledged that Sora may struggle with complex physics simulations and spatial relationships.

The release excludes users in the European Union and United Kingdom, with OpenAI citing ongoing work to meet regional regulatory requirements. Access is also restricted in several other jurisdictions.

Sora was first announced in February 2024 with demonstration videos showing extended sequences and camera movement capabilities. The public launch version operates under tighter constraints than early previews suggested, with shorter maximum durations and resolution limits for most users.

The model represents OpenAI's entry into commercial video generation, a category where competitors including Runway, Pika, and Google have deployed similar systems. OpenAI has not disclosed training data sources or licensing arrangements for video content used in development.