SEEDANCE 2.5

The latest AI video model for 30-second videos and up to 50 multimodal references.