Sora

Sora is an AI model by OpenAI that creates realistic and imaginative scenes from text instructions. Using a diffusion-plus-transformer architecture similar to GPT, it understands and simulates the physical world in motion, generating videos up to one minute long while preserving visual quality and prompt adherence. Sora can also animate still images or extend and fill missing frames in existing videos, serving as a foundation toward world-simulating AGI.