The crew in pipelines.py has three tasks: research, analysis and brief. Use this simplified cost model (an assumption for this question, not CrewAI's exact internals): • Research and analysis each take 2 worker LLM calls (a tool call, then a final answer); the brief takes 1. • Sequential process: only the worker calls. • Hierarchical process: the same worker calls, plus a manager that makes 1 planning call and then, for each of the 3 tasks, 1 call to delegate it and 1 call to review the result. How many LLM calls does the hierarchical run make in total? Give a whole number.