OpenAI unveils Sora, its new text-to-video AI model

OpenAI has launched Sora, its latest text-to-video AI model, capable of transforming text prompts into dynamic scenes with multiple characters and specific motions.

Sora can generate photorealistic videos up to a minute long based on user-written instructions.

According to OpenAI, Sora can understand physical object interactions, interpret props, and create characters with vibrant emotions, enabling it to generate complex and detailed scenes. While it may face challenges in accurately simulating physics, the results are impressive.

Previously, text-to-image generators dominated the field, but advancements in video generation have accelerated. Competitors like Google's Lumiere offer similar text-to-video capabilities.

Currently, Sora is accessible to "red teamers" for risk assessment and select visual artists for feedback.