Hi everyone, what would be the ideal prompt length token when it comes to assistant? For example for models like 4o mini cluster, or 4.1