Wan 2.5 is the widest of the model families this workspace exposes: it is the one version here that offers text to video, image to video, text to image, image editing and video extension from the same family. If a job moves between generating a frame, animating it and then extending the result, staying inside one family means the look is less likely to shift between steps.
The routes offered for Wan 2.5 here are text-to-video, image-to-video, text-to-image, image-edit and video-extend. Each is a separate route in the Wan workspace with its own tab, and each carries its own credit cost.
Model names and available settings are shown live in the form. Availability can change without this page changing, so the live controls are the source of truth — treat the list above as the shape of the family, not a promise about any particular day.
Related pages: text to video, image to video, text to image, image edit, video extend.
Every route here renders a first attempt, not a final asset. Prompts that name the subject, the action and the camera move behave better than prompts that only name a subject, and a fixed seed is the only way to compare two prompts fairly. Check the result before relying on it: a model route can produce a good clip from one prompt and a poor one from its near neighbour.
Cost varies by route, model, resolution and duration. The credit cost for the settings you have chosen is shown beside the form before you submit, and credit prices are on pricing.
Is Wan 2.5 better than the other families here? That depends on the route and the job, not on the version number. Compare with a fixed seed and the same prompt, and judge the output.
Which resolution should I use? Start lower while you are iterating on the prompt, then re-run at the size you actually need to publish.
Where do I start? Open the Wan workspace and pick the tab for the route you need.