OpenAI has introduced Sora 2, a new video and audio generation model, which the company claims is more physically accurate, realistic, and controllable than prior systems. According to the company, the new model can generate scenes such as Olympic-level gymnastics, paddleboard stunts, and skating moves while still respecting physical constraints.
Previous video models often bent reality to satisfy a prompt. For instance, if a basketball player missed a shot, the ball might unrealistically teleport to the hoop. OpenAI says that with Sora 2, the ball would instead bounce off the backboard.
“The model is also a big leap forward in controllability, able to follow intricate instructions spanning multiple shots while accurately persisting world state. It excels at realistic, cinematic, and anime styles,” OpenAI said in a blogpost.
Also read: ChatGPT now lets you buy products directly through chat with Instant Checkout: How it works
Sora 2 is also designed to generate audio at the same time as video. According to the company, the model can create sophisticated background soundscapes, speech, and sound effects with a high degree of realism.
Another feature OpenAI highlights is the ability to “inject” real-world elements. OpenAI claims that by analysing a video of a person, the model can place that individual (appearance and voice included) into any generated scene.
Also read: WhatsApp now supports Live and Motion photos sharing, gets other new features
Alongside the model, OpenAI has also announced a new iOS app called Sora, powered by Sora 2. The app functions like a social platform, allowing users to create content, remix others’ videos, browse a personalised feed, and participate in cameos. “With cameos, you can drop yourself straight into any Sora scene with remarkable fidelity after a short one-time video-and-audio recording in the app to verify your identity and capture your likeness,” the company said.
Also read: Xiaomi 14 Civi available with Rs 17,000 discount during Amazon Great Indian Festival 2025