EN Wan 3.0 is Alibaba’s video model and the everyday choice in the bot, at five credits a second for 480p. It is the model to reach for when you have material to work from: it accepts up to ten reference photos and up to five reference videos on a single draft, runs from 2 to 30 seconds, and will generate without a written prompt at all.
What it is good at
References. Nothing else in the catalogue takes ten photos and five clips at once, and Wan will happily build from them with no prose from you. That makes it the model for restyling something you already have rather than describing something you do not.
It also makes the shortest clips available anywhere in the bot. A two second render at 480p is 10 credits, which is a cheap way to test whether a reference set reads the way you hoped before committing to a longer version. Send them to the bot in one message with the prompt as the caption.
The one thing to watch on price
An attached video is billed. Wan charges the reference clip’s own seconds on top of the output’s seconds, so ten seconds in and ten seconds out is charged as twenty. The two together cannot exceed thirty seconds, and the length picker narrows on its own once a video is attached, so the ceiling is enforced rather than left to you.
Photos are different and simpler: attach as many as you like up to the limit and the price does not move at all. The pricing page has the rate for every model and resolution in one table.
Sound
Sound is a switch here, set before you generate, and it costs nothing either way. Turning it off does not make the clip cheaper, so the only reason to do it is that you want silence.