Clinical Knowledge and Reasoning Abilities of AI Large Language Models in Pharmacy: A Comparative Study on the NAPLEX Exam

Objective This study aims to evaluate the capabilities and limitations of three large language models (LLMs) – GPT-3, GPT-4, and Bard, in the field of pharmaceutical sciences by assessing their pharmaceutical reasoning abilities on a sample North American Pharmacist Licensure Examination (NAPLEX). We also analyze the potential impacts of LLMs on pharmaceutical education and practice. Methods A sample NAPLEX exam consisting of 137 multiple-choice questions was obtained from an online source. GPT-3, GPT-4, and Bard were used to answer the questions by inputting them into the LLMs’ user interface. The answers provided by the LLMs were then compared with the answer key. Results GPT-4 exhibited superior performance compared to GPT-3 and Bard, answering 78.8% of the questions correctly. This score was 11% higher than Bard and 27.7% higher than GPT-3. However, when considering questions that required multiple selections, the performance of each LLM decreased significantly. GPT-4, GPT-3, and Bard only correctly answered 53.6%, 13.9%, and 21.4% of these questions, respectively. Conclusion Among the three LLMs evaluated, GPT-4 was the only model capable of passing the NAPLEX exam. Nevertheless, given the continuous evolution of LLMs, it is reasonable to anticipate that future models will effortlessly pass the exam. This highlights the significant potential of LLMs to impact the pharmaceutical field. Hence, we must evaluate both the positive and negative implications associated with the integration of LLMs in pharmaceutical education and practice.