OpenAI Unveils Sora 2 With Realistic Video and Social App

OpenAI has announced the release of Sora 2 its most advanced video and audio generation model yet. Building on the original Sora launched in early 2024, the new system marks a major leap forward in creating realistic, physically accurate, and controllable video content. Unlike earlier models that often distorted objects or broke the rules of reality to match prompts, Sora 2 can handle complex actions like gymnastic flips, figure skating routines, or paddleboard stunts while maintaining accurate physics, even when characters make mistakes.

 

A key innovation in Sora 2 is its ability to generate synchronized sound effects, dialogue, and background audio alongside video. It supports multiple styles, including cinematic and anime, and can follow detailed prompts across several shots while preserving continuity. Another breakthrough is the option to insert real people, animals, or objects into generated scenes by analyzing a short video recording, enabling lifelike cameos with accurate visuals and voices.

 

To showcase these features, OpenAI is launching a new iOS app simply called Sora. The app allows users to create and remix videos, discover community content, and add themselves or friends into scenes using the Cameo feature. The company believes this represents a new form of communication and creativity, evolving beyond text or emojis. Early testers within OpenAI reported that the feature encouraged new social interactions and playful collaborations.

 

OpenAI says it has designed the app with user well-being in mind. The feed prioritizes creative inspiration rather than maximizing screen time, with parental controls, teen protections, and moderation systems in place. Users retain full control of their likeness, with options to manage or revoke cameo use at any time. The company emphasizes that the platform is built for fun, creativity, and connection rather than addictive scrolling.

 

Sora 2 is launching first in the U.S. and Canada, available through the new app and sora.com. The model will initially be free with generous usage limits, while ChatGPT Pro subscribers will gain access to an upgraded Sora 2 Pro version. OpenAI also plans to release Sora 2 via API in the near future.

 

According to the Sora team, the model demonstrates how scaling video AI brings society closer to building powerful world simulators, with the potential to reshape creativity, communication, and future technology.