Sam Altman, chief executive officer of OpenAI Inc., during a media tour of the Stargate AI data center in Abilene, Texas, US, on Tuesday, Sept. 23, 2025.
Kyle Grillot | Bloomberg | Getty Images
OpenAI is debuting what it bills as its most advanced video generation model yet, alongside a new iOS app designed to bring artificial intelligence-powered video creation to everyday users.
The Sora 2.0 model builds on an earlier version released this year as a research preview. OpenAI said the new release represents a major leap in physical realism, audio-video synchronization, and multi-shot storytelling.
The new app, which is also called Sora, is invite-only to start with and will let users create, remix, and cameo in AI-generated videos using text or images with built-in controls for safety, likeness, and provenance.
The rollout is already raising concerns around intellectual property rules.
The Wall Street Journal reported Monday that OpenAI is notifying studios and talent agencies that, unless they explicitly opt out, their copyrighted material may be reflected in Sora-generated content.
OpenAI says its approach builds on existing content policies used for its image generation tools.
The company views this style of engagement, where users remix or reference familiar fictional worlds, as a form of fan expression and creative play.
Executives said they’ve seen strong user interest in using AI to interact with stories and characters they love, and they believe Sora introduces new ways for people to engage with those fictional universes.
OpenAI has also drawn a clear line between copyrighted IP and individual likeness.
While characters or visual styles from popular franchises may appear unless blocked, individuals cannot be generated at all — via prompt or image — unless they’ve submitted a verified cameo.
The policy is intended to give people full control over how, when, and whether they appear in any AI-generated content.
The app rolls out Tuesday in the U.S. and Canada with an invite-only system. A Pro version, web access, and developer API are all coming soon.
WATCH: Anthropic launches Claude Sonnet 4.5, its latest AI model
(Source)