We're teaching AI to understand and simulate the physical world in motion, with the goal of training models that help people solve problems that require real-world interaction.

Introducing Sora, our text-to-video model. Sora can generate videos up to a minute long while maintaining visual quality and adherence to the user's prompt.