Meta introduced a series of AI models called “Movie Gen” on October 4. These models can produce photorealistic movies up to 16 seconds long, and include sound effects and background music.

Movie Gen isn’t the first multimodal AI model that can generate video and audio from simple text commands, but it does showcase its cutting-edge capabilities. The researchers claim that the application outperformed rival systems in human tests.

According to Meta’s blog post , Movie Gen can currently produce movies at 16 FPS. This is pretty close to professional-quality film footage, considering that pre-digital Hollywood movies were traditionally shot at 24 FPS.

Movie Gen models can generate entirely new movies or modify existing images and videos based on simple text commands. Its most advanced contribution is the ability to produce up to 45 seconds of audio synchronized with motion.

Meta is keeping Movie Gen’s core models under wraps for now, and hasn’t given a timeframe for the product’s release. The company says more security testing will be needed before it can be released.

What do you think of this development? Let us know in the comments.