OpenAI has unveiled Sora 2, its latest video and audio generation model, designed to compete with top-tier platforms like Google’s Veo 3. Available as a free iOS app in the U.S. and Canada, Sora 2 is also accessible to ChatGPT Pro users on sora.com. The model introduces the innovative “cameos” feature, allowing users to insert themselves into AI-generated videos, driving social engagement. OpenAI has implemented an invite-based system for app access, enabling users to control cameo permissions and revoke access as needed. To enhance safety, default limits on video generations for teens and increased moderation against bullying have been established. Sora 2 excels in creating realistic soundscapes, speech, and cinematic visuals, showing improvements in adhering to physical laws compared to previous models. OpenAI plans to release the Sora 2 API, while continuing to support Sora 1 Turbo, ensuring all generated content remains in the existing library.
Source link
Share
Read more