EUVIMEDEUVIMED entdecken
BeobachtungsstudieVergleichsstudieOpen Access · CC BY 4.0

AI performance in emergency medicine fellowship examination: comparative analysis of Chatgpt-4o, Gemini 2.0, Claude 3.5, and Deepseek R1 models

İshak Şan iD, Medine Akkan Öz iD, Mehmet Yortanli iD, Murat Genç iD, Bensu Bulut iD, Ayşenur Gür iD, Hüseyin Mutlu iD, Ramiz Yazici iD, Mustafa Önder Gönen iD

Turkish Journal of Medical Sciences · 14.10.2025 · 55 (5) · 1292-1299

DOI: 10.55730/1300-0144.6083 ↗

Kurz gefasst

Beobachtungsstudie in der Zeitschrift Turkish Journal of Medical Sciences, online erschienen am 14.10.2025. Verfasst von İshak Şan und 8 weiteren Autor:innen. Beteiligte Einrichtung: Department of Emergency Medicine, Health Science University, Bilkent City Hospital, Ankara, Turkiye. Hauptthemen laut MEDLINE: Emergency Medicine, Educational Measurement, Fellowships and Scholarships. Die Arbeit wurde bisher 7-mal zitiert (Crossref). Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Background/aim: This study evaluated the accuracy rates and response consistency of four different large language models (ChatGPT- 4o, Gemini 2.0, Claude 3.5, and DeepSeek R1) in answering questions from the Emergency Medicine Fellowship Examination (YDUS), which was administered for the first time in Türkiye. Materials and methods: In this observational study, 60 multiple-choice questions from the Emergency Medicine YDUS administered on 15 December 2024, were classified as knowledge-based (n = 26), visual content (n = 2), and case-based (n = 32). Each question was presented three times to the four large language models. The models’ accuracy rates were evaluated according to overall accuracy, strict accuracy, and ideal accuracy criteria. Response consistency was measured using Fleiss’ Kappa test.

Results

The ChatGPT-4o model was the most successful in terms of overall accuracy (90.0%), while DeepSeek R1 showed the lowest performance (76.7%). Claude 3.5 (83.3%) and Gemini 2.0 (80.0%) demonstrated moderate success. When analyzed by category, ChatGPT-4o achieved the highest success with 92.3% accuracy in knowledge-based questions and 90.6% in case-based questions. In terms of response consistency, the Claude 3.5 model (Fleiss’ Kappa = 0.68) showed the highest consistency, while Gemini 2.0 (Fleiss’ Kappa = 0.49) showed the lowest. Inconsistent hallucinations were more frequent in the Gemini 2.0 and DeepSeek R1 models, whereas persistent hallucinations were less common in the ChatGPT-4o and Claude 3.5 models.

Conclusion

Large language models can achieve high accuracy rates for knowledge and clinical reasoning questions in emergency medicine but show differences in terms of response consistency and hallucination tendency. While these models have significant potential for use in medical education and as clinical decision support systems (CDSS), they need further development to provide reliable, up-to-date, and accurate information.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Emergency MedicineEducational MeasurementFellowships and ScholarshipsHumansTurkeyClinical CompetenceGenerative Artificial Intelligence

Stichwörter der Autor:innen

Emergency medicineartificial intelligencefellowshiplanguage

Autor:innen und Institutionen

  • Department of Emergency Medicine, Health Science University, Bilkent City Hospital, Ankara, Turkiye.
  • Department of Emergency Medicine, Gülhane Training and Research Hospital, Ankara, Turkiye.
  • Department of Emergency Medicine, Konya Numune Hospital, Konya, Turkiye.
  • Department of Emergency Medicine, Ankara Training and Research Hospital, Ankara, Turkiye.
  • Department of Emergency Medicine, Health Science University, Gülhane Training and Research Hospital, Ankara, Turkiye.
  • Department of Emergency Medicine, Etimesgut State Hospital, Ankara, Turkiye.

ORCID-Profile: İshak Şan, Medine Akkan Öz, Mehmet Yortanli, Murat Genç, Bensu Bulut, Ayşenur Gür, Hüseyin Mutlu, Ramiz Yazici, Mustafa Önder Gönen

Zitieren

APA

Şan, İ., Akkan Öz, M., Yortanli, M., Genç, M., Bulut, B., Gür, A., Mutlu, H., Yazici, R., & Gönen, M. Ö. (2025). AI performance in emergency medicine fellowship examination: comparative analysis of Chatgpt-4o, Gemini 2.0, Claude 3.5, and Deepseek R1 models. Turkish Journal of Medical Sciences, 55(5), 1292-1299. https://doi.org/10.55730/1300-0144.6083

Vancouver

Şan İ, Akkan Öz M, Yortanli M, Genç M, Bulut B, Gür A, et al. AI performance in emergency medicine fellowship examination: comparative analysis of Chatgpt-4o, Gemini 2.0, Claude 3.5, and Deepseek R1 models. Turkish Journal of Medical Sciences. 2025;55(5):1292-1299. doi:10.55730/1300-0144.6083

BibTeX
@article{an2025,
  author = {Şan, İshak and Akkan Öz, Medine and Yortanli, Mehmet and Genç, Murat and Bulut, Bensu and Gür, Ayşenur and Mutlu, Hüseyin and Yazici, Ramiz and Gönen, Mustafa Önder},
  title = {AI performance in emergency medicine fellowship examination: comparative analysis of Chatgpt-4o, Gemini 2.0, Claude 3.5, and Deepseek R1 models},
  journal = {Turkish Journal of Medical Sciences},
  year = {2025},
  volume = {55},
  number = {5},
  pages = {1292--1299},
  doi = {10.55730/1300-0144.6083}
}
Kostenlos starten →