OpenAI has launched Sora 2, an advanced text-to-video model competing with Google's Veo-3. Sora 2 improves realism and controllability, allowing users to create and remix videos.