A multi-model longitudinal assessment of ChatGPT performance on medical residency examinations

IntroductionChatGPT, a generative artificial intelligence, has potential applications in numerous fields, including medical education. This potential can be assessed through its performance on medical exams. Medical residency exams, critical for entering medical specialties, serve as a valuable benc...

Full description

Saved in:
Bibliographic Details
Main Authors: Maria Eduarda Varela Cavalcanti Souto, Alexandre Chaves Fernandes, Ana Beatriz Santana Silva, Louise Helena de Freitas Ribeiro, Thales Allyrio Araújo de Medeiros Fernandes
Format: Article
Language:English
Published: Frontiers Media S.A. 2025-08-01
Series:Frontiers in Artificial Intelligence
Subjects:
Online Access:https://www.frontiersin.org/articles/10.3389/frai.2025.1614874/full
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849229200721969152
author Maria Eduarda Varela Cavalcanti Souto
Alexandre Chaves Fernandes
Ana Beatriz Santana Silva
Louise Helena de Freitas Ribeiro
Thales Allyrio Araújo de Medeiros Fernandes
author_facet Maria Eduarda Varela Cavalcanti Souto
Alexandre Chaves Fernandes
Ana Beatriz Santana Silva
Louise Helena de Freitas Ribeiro
Thales Allyrio Araújo de Medeiros Fernandes
author_sort Maria Eduarda Varela Cavalcanti Souto
collection DOAJ
description IntroductionChatGPT, a generative artificial intelligence, has potential applications in numerous fields, including medical education. This potential can be assessed through its performance on medical exams. Medical residency exams, critical for entering medical specialties, serve as a valuable benchmark.Materials and methodsThis study aimed to assess the accuracy of ChatGPT-4 and GPT-4o in responding to 1,041 medical residency questions from Brazil, examining overall accuracy and performance across different medical areas, based on evaluations conducted in 2023 and 2024. The questions were classified into higher and lower cognitive levels according to Bloom’s taxonomy. Additionally, questions answered incorrectly by both models were tested using the recent GPT models that use chain-of-thought reasoning (e.g., o1-preview, o3, o4-mini-high) with evaluations carried out in both 2024 and 2025.ResultsGPT-4 achieved 81.27% accuracy (95% CI: 78.89–83.64%), while GPT-4o reached 85.88% (95% CI: 83.76–88.00%), significantly outperforming GPT-4 (p < 0.05). Both models showed reduced accuracy on higher-order thinking questions. On questions that both models failed, GPT o1-preview achieved 53.26% accuracy (95% CI: 42.87–63.65%), GPT o3 47.83% (95% CI: 37.42–58.23%) and o4-mini-high 35.87% (95% CI: 25.88–45.86%), with all three models performing better on higher-order questions.ConclusionArtificial intelligence could be a beneficial tool in medical education, enhancing residency exam preparation, helping to understand complex topics, and improving teaching strategies. However, careful use of artificial intelligence is essential due to ethical concerns and potential limitations in both educational and clinical practice.
format Article
id doaj-art-e37dc193c6b147f49b60b46a282976ec
institution Kabale University
issn 2624-8212
language English
publishDate 2025-08-01
publisher Frontiers Media S.A.
record_format Article
series Frontiers in Artificial Intelligence
spelling doaj-art-e37dc193c6b147f49b60b46a282976ec2025-08-22T05:26:44ZengFrontiers Media S.A.Frontiers in Artificial Intelligence2624-82122025-08-01810.3389/frai.2025.16148741614874A multi-model longitudinal assessment of ChatGPT performance on medical residency examinationsMaria Eduarda Varela Cavalcanti Souto0Alexandre Chaves Fernandes1Ana Beatriz Santana Silva2Louise Helena de Freitas Ribeiro3Thales Allyrio Araújo de Medeiros Fernandes4Department of Biomedical Sciences, School of Health Sciences, State University of Rio Grande do Norte, Mossoró, BrazilInstitute of Mathematics and Computer Sciences, University of São Paulo, São Paulo, BrazilDepartment of Biomedical Sciences, School of Health Sciences, State University of Rio Grande do Norte, Mossoró, BrazilDepartment of Biomedical Sciences, School of Health Sciences, State University of Rio Grande do Norte, Mossoró, BrazilDepartment of Biomedical Sciences, School of Health Sciences, State University of Rio Grande do Norte, Mossoró, BrazilIntroductionChatGPT, a generative artificial intelligence, has potential applications in numerous fields, including medical education. This potential can be assessed through its performance on medical exams. Medical residency exams, critical for entering medical specialties, serve as a valuable benchmark.Materials and methodsThis study aimed to assess the accuracy of ChatGPT-4 and GPT-4o in responding to 1,041 medical residency questions from Brazil, examining overall accuracy and performance across different medical areas, based on evaluations conducted in 2023 and 2024. The questions were classified into higher and lower cognitive levels according to Bloom’s taxonomy. Additionally, questions answered incorrectly by both models were tested using the recent GPT models that use chain-of-thought reasoning (e.g., o1-preview, o3, o4-mini-high) with evaluations carried out in both 2024 and 2025.ResultsGPT-4 achieved 81.27% accuracy (95% CI: 78.89–83.64%), while GPT-4o reached 85.88% (95% CI: 83.76–88.00%), significantly outperforming GPT-4 (p < 0.05). Both models showed reduced accuracy on higher-order thinking questions. On questions that both models failed, GPT o1-preview achieved 53.26% accuracy (95% CI: 42.87–63.65%), GPT o3 47.83% (95% CI: 37.42–58.23%) and o4-mini-high 35.87% (95% CI: 25.88–45.86%), with all three models performing better on higher-order questions.ConclusionArtificial intelligence could be a beneficial tool in medical education, enhancing residency exam preparation, helping to understand complex topics, and improving teaching strategies. However, careful use of artificial intelligence is essential due to ethical concerns and potential limitations in both educational and clinical practice.https://www.frontiersin.org/articles/10.3389/frai.2025.1614874/fullgenerative artificial intelligencemedical residency examinationsmedical educationartificial intelligencechain-of-thought reasoninglarge language model
spellingShingle Maria Eduarda Varela Cavalcanti Souto
Alexandre Chaves Fernandes
Ana Beatriz Santana Silva
Louise Helena de Freitas Ribeiro
Thales Allyrio Araújo de Medeiros Fernandes
A multi-model longitudinal assessment of ChatGPT performance on medical residency examinations
Frontiers in Artificial Intelligence
generative artificial intelligence
medical residency examinations
medical education
artificial intelligence
chain-of-thought reasoning
large language model
title A multi-model longitudinal assessment of ChatGPT performance on medical residency examinations
title_full A multi-model longitudinal assessment of ChatGPT performance on medical residency examinations
title_fullStr A multi-model longitudinal assessment of ChatGPT performance on medical residency examinations
title_full_unstemmed A multi-model longitudinal assessment of ChatGPT performance on medical residency examinations
title_short A multi-model longitudinal assessment of ChatGPT performance on medical residency examinations
title_sort multi model longitudinal assessment of chatgpt performance on medical residency examinations
topic generative artificial intelligence
medical residency examinations
medical education
artificial intelligence
chain-of-thought reasoning
large language model
url https://www.frontiersin.org/articles/10.3389/frai.2025.1614874/full
work_keys_str_mv AT mariaeduardavarelacavalcantisouto amultimodellongitudinalassessmentofchatgptperformanceonmedicalresidencyexaminations
AT alexandrechavesfernandes amultimodellongitudinalassessmentofchatgptperformanceonmedicalresidencyexaminations
AT anabeatrizsantanasilva amultimodellongitudinalassessmentofchatgptperformanceonmedicalresidencyexaminations
AT louisehelenadefreitasribeiro amultimodellongitudinalassessmentofchatgptperformanceonmedicalresidencyexaminations
AT thalesallyrioaraujodemedeirosfernandes amultimodellongitudinalassessmentofchatgptperformanceonmedicalresidencyexaminations
AT mariaeduardavarelacavalcantisouto multimodellongitudinalassessmentofchatgptperformanceonmedicalresidencyexaminations
AT alexandrechavesfernandes multimodellongitudinalassessmentofchatgptperformanceonmedicalresidencyexaminations
AT anabeatrizsantanasilva multimodellongitudinalassessmentofchatgptperformanceonmedicalresidencyexaminations
AT louisehelenadefreitasribeiro multimodellongitudinalassessmentofchatgptperformanceonmedicalresidencyexaminations
AT thalesallyrioaraujodemedeirosfernandes multimodellongitudinalassessmentofchatgptperformanceonmedicalresidencyexaminations