Background: Recent advancements in artificial intelligence (AI), particularly in large language models (LLMs), have started a new era of innovation across various fields, with medicine at the forefront of this technological revolution. Many studies indicated that at the current level of development, LLMs can pass different board exams. However, the ability to answer specific subject-related questions requires validation. Objective: The objective of this study was to conduct a comprehensive analysis comparing the performance of advanced LLM chatbots—Claude (Anthropic), GPT-4 (OpenAI), Gemini (Google), and Copilot (Microsoft)—against the academic results of medical students in the medical biochemistry course. Methods: We used 200 USMLE (United States Medical Licensing Examination)–style multiple-choice questions (MCQs) selected from the course exam database. They encompassed various complexity levels and were distributed across 23 distinctive topics. The questions with tables and images were not included in the study. The results of 5 successive attempts by Claude 3.5 Sonnet, GPT-4‐1106, Gemini 1.5 Flash, and Copilot to answer this questionnaire set were evaluated based on accuracy in August 2024. Statistica 13.5.0.17 (TIBCO Software Inc) was used to analyze the data’s basic statistics. Considering the binary nature of the data, the chi-square test was used to compare results among the different chatbots, with a statistical significance level of P<.05. Results: On average, the selected chatbots correctly answered 81.1% (SD 12.8%) of the questions, surpassing the students’ performance by 8.3% (P=.02). In this study, Claude showed the best performance in biochemistry MCQs, correctly answering 92.5% (185/200) of questions, followed by GPT-4 (170/200, 85%), Gemini (157/200, 78.5%), and Copilot (128/200, 64%). The chatbots demonstrated the best results in the following 4 topics: eicosanoids (mean 100%, SD 0%), bioenergetics and electron transport chain (mean 96.4%, SD 7.2%), hexose monophosphate pathway (mean 91.7%, SD 16.7%), and ketone bodies (mean 93.8%, SD 12.5%). The Pearson chi-square test indicated a statistically significant association between the answers of all 4 chatbots (P<.001 to P<.04). Conclusions: Our study suggests that different AI models may have unique strengths in specific medical fields, which could be leveraged for targeted support in biochemistry courses. This performance highlights the potential of AI in medical education and assessment.

BACKGROUND Recent advancements in artificial intelligence (AI), particularly in large language models (LLMs), have started a new era of innovation across various fields, with medicine at the forefront of this technological revolution. Many studies indicated that at the current level of development, LLMs can pass different board exams. However, the ability to answer specific subject-related questions requires validation. OBJECTIVE The objective of this study was to conduct a comprehensive analysis comparing the performance of advanced LLM chatbots - Claude (Anthropic), GPT-4 (OpenAI), Gemini (Google), and Copilot (Microsoft), against the academic results of medical students in the medical biochemistry course. METHODS We used 200 USMLE-style multiple-choice questions selected from the course exam database. They encompassed various complexity levels and were distributed across 23 distinctive topics. The questions with tables and images were not included in the study. The results of 5 successive attempts by Claude 3.5 Sonnet, GPT-4-1106, Gemini 1.5 Flash, and Copilot to answer this questionnaire set were evaluated based on accuracy in August 2024. Statistica 13.5.0.17 (TIBC® Statistica™) was used to analyze the data's basic statistics. Considering the binary nature of the data, the Chi-square test was utilized to compare results among the different chatbots, with a statistical significance level of P<.05. RESULTS On average, the selected chatbots correctly answered 81.1±12.8% of the questions, surpassing the students' performance by 8.3% (P=.017). In this study, Claude showed the best performance in biochemistry MCQs, correctly answering 92.5% of questions, followed by GPT-4 (85.1%), Gemini (78.5%), and Copilot (64%). The chatbots demonstrated the best results in the following four topics: Eicosanoids (100%), Bioenergetics and Electron transport chain (96.4±7.2), Hexose monophosphate pathway (91.7±16.7), and Ketone bodies (93.8±12.5). The Pearson Chi-square test indicated a statistically significant association between the answers of all 4 chatbots (P<.001- P<.044). CONCLUSIONS Our study suggests that different AI models may have unique strengths in specific medical fields, which could be leveraged for targeted educational support in biochemistry courses. This performance highlights the potential of AI in medical education and assessment.

Dyslipidemia driven by excess triglyceride-rich lipoproteins (TGRLs) remains a major modifiable determinant of atherosclerotic cardiovascular disease (ASCVD). Over the past decade, the angiopoietin-like protein (ANGPTL) family (specifically ANGPTL3, ANGPTL4, and ANGPTL8) has emerged as a central regulatory axis controlling lipoprotein lipase (LPL) activity and the tissue-specific partitioning of circulating triglycerides. Human genetic studies of loss-of-function variants and the clinical success of biologic ANGPTL3 inhibitors have validated this axis as a therapeutic target, but these agents are costly, parenterally administered, and reserved for severe or refractory dyslipidemia. In parallel, a growing body of preclinical work has identified naturally occurring bioactive compounds, including citrus polymethoxyflavones, dietary polyphenols, marine- and plant-derived fatty acids, royal jelly constituents, and polysaccharides, which can modulate ANGPTL3, ANGPTL4, or ANGPTL8 expression and downstream LPL activity. This review synthesizes the molecular mechanisms by which these compounds engage the ANGPTL3/4/8 axis, situates the supporting evidence within a hierarchy that spans hepatocyte and myocyte cell culture, rodent and non-human primate models, and the scarce available human trials, and contrasts this evidence base against the mature clinical data underpinning pharmacological ANGPTL inhibitors such as evinacumab. We identify liver X receptor alpha (LXRα) and hepatocyte nuclear factor 4 alpha (HNF4α) as proposed transcriptional nodes for structurally diverse natural ligands, noting that no compound has been shown to engage either receptor directly, and that a requirement for the transcription factor has been established by loss-of-function experiment only for HNF4α, in a single compound and cell line, and we discuss translational obstacles including limited oral bioavailability, interspecies dosing discrepancies, and the near-total absence of adequately powered human outcome trials. We conclude that natural bioactive compounds represent a mechanistically plausible but clinically immature adjunctive strategy for modulating the ANGPTL3/4/8 axis, and we outline the preclinical and clinical research priorities required to move this field toward evidence-based application in cardiometabolic disease management.