Sora 2 is OpenAI’s flagship video and audio generation model that advances beyond the original Sora with more accurate physics, sharper realism, synchronized dialogue and sound effects, and far greater control over multi‑shot sequences while preserving world state. It powers the new Sora app for collaborative creation (including remixing and opt‑in ‘cameos’ of yourself or friends) and is initially rolling out in the U.S. and Canada, with Sora 2 Pro available to ChatGPT Pro users on sora.com and API access planned. OpenAI’s Sora 2 System Card outlines rigorous red‑teaming and safety mitigations—particularly around likeness consent and misleading content—reflecting an iterative, responsibility‑first deployment approach.
