>>59443676>>59444052Ah it seems it's not the motion model, the prompt token limit is still in effect just like in the client version.
Simple prompts I can get the 16 frames but if I go over it does that split. At least from the couple of test I ran.
Are your prompts running over the 75 tokens?