OpenAI's Sora 2 video generation model. Features significant improvements in video quality, physics simulation, and prompt adherence compared to the original Sora. Supports longer duration, higher resolution, and more complex scene generation. Approximate launch: 2025.
