Meta Movie Gen looks incredible — 5 impressive examples

Text-to-video offerings within AI are growing quickly, and it feels like plenty of models offer the option to request complex video generation from a text prompt.

Meta Movie Gen was only unveiled earlier this month, but 'the company that Facebook built' has more resources than many of its rivals including the ability to add sound.

Like OpenAI's SORA or Google's Veo it isn't currently available for the public to use, restricted to the hand-picked few. But that has given us an insight into what will be possible once the full model is eventually released.

With that in mind, we've done some digging to find the most impressive examples of Meta Movie Gen so far, and there are plenty to choose from.

Snake in a forest

I'll level with you, snakes terrify me, but it's hard not to be impressed as this AI one slithers through a series of vines and roots.

The addition of a soundtrack prompt seemingly suggests tension, too, meaning the underlying model may agree that they're unsettling, but the way it moves behind leaves is very impressive from a processing point of view.

Jellyfish everywhere

This Instagram post showcases changes made to existing footage, but stick with it and you'll see an incredible look at an underwater scene with literally hundreds of jellyfish underwater.

Most impressive is the way the light from the jellyfish bounces off of the subject's face and body as they swim upward.

Skateboarding Sounds

Sure, this is a cool-looking shot, but it's the sound that's perhaps most impressive. Watch as the skateboarder moves across the concrete, creating that unmistakable sound of skateboard wheels on skate park concrete.

The sound of takeoff and landing are just right, too, and the way the camera follows the action is really close to what you'd see in, say, a commercial.

Soup lady

This one falls right into 'Uncanny Valley' for me, but the detail on this subject's face is wild - look at the wrinkles, her subtle smile, and the way her hair curls in the fringe.

I'm not sure the soup she's eating looks all that appealing, but it's still very impressive how the mannerisms of her eating it look just about human enough to pass.

Child running on beach

This was one that Meta used to demonstrate the new model, and it's not hard to see why. Aside from the wind, the footprints, and the kite, I'm most taken aback by the actual locomotion of the child subject.

As a parent, I know that my son (who is of a similar age to this subject), runs a certain way. It's a little clumsy, almost like there's every chance he's going to fall at any time. Impressively, I'd say this just about captures that same fragility of stride.

More from Tom's Guide

• Google Gemini — 5 tips to get better results with your prompts
• Apple is '2 years behind' on AI but here's why I think it will still end up ahead
• I used Suno AI's new Suno Scenes feature to turn a family photo into a song — here's how it works