Transformer-based large language models are gaining a foothold in the field of machine learning research, with a plethora of applications in natural language processing, biology, chemistry, and computer programming. These models extreme scaling and reinforcement learning capabilities have led to improvements in the quality of generated text, enabling machines to perform various tasks and reason about their choices. In this paper, we present an intelligent agent system that uses multiple large language models to autonomously design, plan, and execute scientific experiments. Our agents capabilities are demonstrated through three distinct examples, including successful performance in catalyzed cross-coupling reactions. The safety implications of using such systems are discussed, and measures to prevent misuse are proposed. The paper also discusses various glossary terms such as LLMs, Agents, and Prompt-providers, which are essential to understanding the papers content.