The monetary policy is a crucial tool for guiding national economic regulations, as it upholds economic stability and fosters economic development. Specifically, the policies implemented by the Chinese Central Bank play a vital role in managing inflation and bolstering economic growth. These policies directly influence the living standards of the population and their financial decisions. However, in the past, interpreting these policies required a strong professional background in economics or finance, along with significant manual effort. The emergence of Q&A systems has provided the general public with better access to understanding the measures taken by the central bank and accessing monetary policy information more efficiently. Unfortunately, most of these Q&A systems employ basic and outdated natural language processing (NLP) algorithms and models, resulting in limited abilities to understand and reason about the policies and questions, ultimately leading to poor outcomes. In this paper, we present our approach to creating and training a state-of-the-art Q&A system for Chinese Central Bank policies using large language models (LLMs). Specifically, we demonstrate three of the most popular Chinese-based LLMs: IFlytek, ChatGPT, and LangChain-ChatGLM. Among them, IFlytek and ChatGPT are cloud-service-based LLMs trained through the prompt-tuning method. On the other hand, LangChain-ChatGLM utilizes the Langchain framework, which includes autoregressive gap filling and combines the advantages of autocoding and autoregressive pretraining. The ChatGLM-6B model is used for knowledge learning in this approach. We compare the performance of all these LLMs and publish the comparison results for future study.