Core Concepts
OpenAI introduces Sora, a text-to-video model, to enhance AI's understanding of the physical world through motion simulation.
Abstract
OpenAI has unveiled Sora, a new text-to-video model named after the Japanese word for "sky." This innovative model, powered by a transformer architecture like GPT models, can create high-quality videos up to a minute long based on user input. Notably, Sora can animate still images accurately, bringing them to life with intricate details and fidelity.
Stats
Sora can generate videos lasting up to a minute.
The model maintains high visual quality and fidelity to user prompts.
It can animate still images accurately and capture even the smallest details.