Close-up, A little girl wearing a red hoodie in winter strikes a match.
River. Water is rushing down a stream and pouring.
Tactical soldier in futuristic body armor holding an advanced rifle.
Woman practicing yoga in a lush garden setting.
Generate a mysterious ghost figure in an atmospheric scene set in an old attic.
Generate a surreal scene of a dog melting while sitting on a cozy living room floor.
Man sitting at an outdoor café in a vintage European cityscape during golden hour.
Wide shot: A caravan of camels winds its way through the endless golden dunes.
Whimsical and vibrant scene of a panda riding a bicycle through a bustling city street.
Playful and humorous scene of a fluffy cat curiously sniffing and nibbling on a large burger.
In the gym, a woman in workout clothes runs on a treadmill.
Powerful off-road truck speeding through a dirt track at sunset
The largest open-source video generation model available today, delivering superior quality and performance.
Advanced 3D VAE architecture ensures smooth, natural motion and exceptional visual consistency throughout videos.
Superior text understanding capabilities for better text-to-video alignment and more accurate results.
Complete access to code and model weights on GitHub, enabling community contribution and improvements.
Outperforms previous state-of-the-art models with 68.5% text alignment and 96.4% visual quality scores.
Supports various video resolutions including 720p×1280p vertical resolution to suit different needs. Only 1280px720p is currently supported in generation. Stay tuned for more!