On the eve of China's National Day holiday, OpenAI unveiled its state-of-the-art video and audio generation model, Sora 2. This innovative model is capable of tackling tasks that have traditionally posed challenges for video generation models. For instance, it can flawlessly simulate athletes' Olympic gymnastics routines, execute backflips on paddleboards, and accurately depict physical properties such as buoyancy and rigidity. Sora 2 shines in its ability to produce content with remarkable realism, whether in cinematic or anime styles. Moreover, for the very first time, it enables the synchronous generation of video and audio, significantly boosting the realism and physical precision of the videos produced. In tandem with this release, OpenAI also rolled out the iOS application Sora, which empowers users to craft and remix AI-generated videos, as well as seamlessly integrate themselves or their friends into these videos through the 'Cameo' functionality.