On Tuesday, OpenAI unveiled Sora 2, its second-generation video-synthesis AI model, capable of generating videos with synchronized dialogue and sound effects—an impressive first for the company. Accompanying this release is a new iOS social app that allows users to create “cameos” in AI-generated videos. A launch video features a photorealistic Sam Altman, OpenAI’s CEO, engaging in various imaginative scenarios. Sora 2 boasts advancements in audio realism, generating sophisticated soundscapes and dialogue. It builds on Google’s Veo 3 and Alibaba’s Wan 2.5 by enhancing visual consistency and enabling complex instruction handling over multiple shots. OpenAI likens Sora 2 to its landmark moment with ChatGPT, showcasing improved physical accuracy for dynamic movements, such as Olympic routines. The model addresses previous shortcomings, ensuring more realistic outcomes, like a basketball rebounding realistically instead of teleporting. Sora 2 positions OpenAI as a leader in AI video generation, setting new industry standards.
Source link
