WAN

WAN is Alibaba's open video model family. It generates short clips from text or images, holds up well on human motion, and costs less per second than most closed models. A common pick when you want lots of takes without burning through credits.

Between 5 and 68 credits per 5 second clip depending on the variant. You buy credits once and spend them on any model, no subscription needed.

The WAN family on BasedLabs covers text to image, image to video, video to video, text to video and image editing across 17 variants. Pick a variant below and it opens in the generator with that model preselected.

Text to video builds the whole scene from your prompt, so it is best when you do not have a starting visual. Image to video animates a picture you already have, which gives you much more control over composition and characters. If you care about how the first frame looks, generate an image first and animate it.

You need a free BasedLabs account, which comes with starter credits. After that you only pay for what you generate.