OpenAI officially unveiled Sora 2, the second version of its video generation model, and simultaneously launched the "Sora" iOS app on October 1, 2025. The new model shows significant improvements: enhanced physical consistency of objects in the frame, better controllability, and, most importantly, the ability to generate synchronized sound that matches the video. The "Sora" app, which began rolling out on an invite-only basis, allows users to create short video clips and use a new "cameo" feature to generate digital likenesses of people. As WIRED notes in its detailed review, OpenAI has implemented strict safety measures: creating "cameos" of public figures is prohibited, and all generated content is watermarked using the C2PA standard to prevent misuse. This launch moves video generation technology from the demo stage to an accessible consumer product category and raises new questions in ethics and safety.
OpenAI Launches Sora 2 and App with Sound Generation and "Cameos"
