This paper introduces the Agent-based Simulated Financial Market (ASFM), a novel framework that uses large language models (LLMs) to simulate financial markets. The ASFM consists of a simulated stock market with a real-order matching system and LLM-based trading agents. Each agent has a profile, observation, and tool-learning-based action module, enabling them to understand market dynamics and financial policies, and make decisions aligned with their trading strategies. The authors validate the ASFM's accuracy by comparing its reactions to real stock market scenarios, including interest rate changes and inflation shocks. They also explore two key economic research areas: the impact of trader behavior bias and the large trader impact. The results show that ASFM consistently aligns with preliminary findings in economics research, demonstrating its potential as a new paradigm for economic modeling. The paper concludes by discussing future directions, such as incorporating more financial market participants and behaviors, and enhancing the complexity of agent behaviors.This paper introduces the Agent-based Simulated Financial Market (ASFM), a novel framework that uses large language models (LLMs) to simulate financial markets. The ASFM consists of a simulated stock market with a real-order matching system and LLM-based trading agents. Each agent has a profile, observation, and tool-learning-based action module, enabling them to understand market dynamics and financial policies, and make decisions aligned with their trading strategies. The authors validate the ASFM's accuracy by comparing its reactions to real stock market scenarios, including interest rate changes and inflation shocks. They also explore two key economic research areas: the impact of trader behavior bias and the large trader impact. The results show that ASFM consistently aligns with preliminary findings in economics research, demonstrating its potential as a new paradigm for economic modeling. The paper concludes by discussing future directions, such as incorporating more financial market participants and behaviors, and enhancing the complexity of agent behaviors.