Meet Sora, our text-to-video model! Sora is designed to help people solve real-world problems by understanding and simulating motion in the physical world. This AI model can generate high-quality, minute-long videos based on user prompts. Today, we’re giving access to red teamers to assess potential risks and visual artists to gather feedback for improvement. It’s important for us to involve external partners and give the public a glimpse into the future of AI capabilities. Sora is capable of creating complex scenes with multiple characters, realistic motion, and accurate details. It can interpret language prompts and bring vibrant emotions to characters. However, Sora has some limitations in accurately simulating physics and understanding cause and effect. It may also struggle with spatial and temporal details.
https://openai.com/sora