OpenAI Wang Fried: The new Sora model generates a 1-minute high-quality video in one sentence

IT Home reported on February 16 that OpenAI recently released the Sora model, which can generate a video content based on the text description entered by the user.

The model can deeply simulate the real physical world, marking a major leap forward in AI's ability to understand and interact with real-world scenarios.

Sora uses the prompts entered by the user to create videos up to a minute long while maintaining high visual quality. For any artist, filmmaker, or student who needs to make a video, the possibilities are endless.

OpenAI says Sora has been tested by Team Red, a group that plays the role of an enemy or competitor in a cybersecurity exercise, to assess potential hazards or risks.

In addition, OpenAI also invited a team of professional creatives to test it to give feedback on its usefulness in a professional environment. Based on this feedback, OpenAI plans to improve Sora to ensure that it effectively meets the needs of its users. The demo video is very surreal.

Sora can create complex scenes with multiple people, specific motion types, and detailed backgrounds. It generates videos that accurately reflect the user's prompts. For example, Sora can make videos of stylish women walking the neon-lit streets of Tokyo, videos of giant mammoths in the snow, and even movie trailers of astronaut adventures.

However, Sora has limitations, including challenges in simulating the physics of complex scenarios and understanding specific causal scenarios. OpenAI says Sora can also confuse spatial details and struggle with accurately describing temporal events.