Wan's advanced AI video model with stronger prompt understanding — generate longer, more complex video scenes with consistent characters, dynamic camera movement, and rich visual detail