Seedance 2.0 Mini is the practical version to try when speed and iteration matter. It keeps the same core creative direction style as Seedance 2.0, but focuses on 480p and 720p output for faster text-to-video and image-to-video work.
Use text-to-video when you want the model to invent the shot from a written prompt. Use image-to-video when the first frame already has the subject, product, character, or composition you want to preserve.
The strongest Mini workflow is reference-first: start with a clear image, describe the movement, and optionally add an end image when the shot needs a specific final frame instead of open-ended motion.