OpenAI launches Sora 2 video generation model and invite-only iOS app

OpenAI on Tuesday released Sora 2, a next generation audio and video generation model, and an accompanying invite only iOS app called Sora that lets users create, remix and share short AI generated videos. The rollout begins in the United States and Canada; OpenAI said the app will initially be invite only and free to use, with ChatGPT Pro subscribers given early access to a higher quality Sora 2 Pro model and broader developer access planned via an API.

OpenAI described Sora 2 as an advance in realism and controllability compared with earlier systems. The company said the model better preserves physical laws and scene continuity, can follow multishot instructions across a sequence, and synchronizes audio with visuals, including dialogue, sound effects and background soundscapes. A core capability of the app is Cameos: users may upload a one time video and audio recording to create an authenticated likeness and voice that can be inserted into generated scenes. Cameos are opt in, can be shared with friends, and can be revoked by their owner, OpenAI said.

The Sora app features a scrolling, personalized feed that OpenAI said is designed to encourage creation rather than passive...