What does this research mean for the field?

Gemini 3.5 Flash and ChatGPT-5 achieve high accuracy rates (90.6% and 88.2%) on dental specialty examinations, significantly outperforming other large language models, though performance varies considerably across different dental disciplines. Novelty: ClaimNovelty.INCREMENTAL. Consensus alignment: ConsensusAlignment.NEUTRAL.

What question did this study set out to answer?

To evaluate and compare the accuracy of five large language models in answering clinical questions from Turkey's dental specialty examination (DUS).

September 5, 2026European Journal Of Dental EducationOpen Access

Evaluating the Accuracy of Large Language Models in Dentistry: A Multi‐Model Study Using Clinical Questions From Turkey's Dental Specialty Exams

View Full Paper

Ask AI

Bookmark

Authors

NKNihan Kaya FŞFatih Şengül PÇPeriş Çelikel

Discussion

Member takes

Overview

Comparative evaluation reveals variable accuracy among five large language models answering dental specialty exam questions, indicating promise as educational aids.

Key Points

To evaluate and compare the accuracy of five large language models in answering clinical questions from Turkey's dental specialty examination (DUS).
Evaluated 1,026 clinical science questions from 13 Turkish dental specialty examinations (2012–2021) across five LLMs: ChatGPT-5, Gemini 3.5 Flash, Microsoft Copilot, DeepSeek-R1, and Grok 4.
Categorized questions by dental specialty and question type (information-based vs. case-based), scoring single-prompt outputs against official answer keys using chi-square and Z tests with Bonferroni correction.
Overall accuracy ranged from 79.2% to 90.6%, with Gemini 3.5 Flash (90.6%) and ChatGPT-5 (88.2%) achieving the highest performance, while Microsoft Copilot, DeepSeek-R1, and Grok 4 were significantly less accurate (p < 0.001).
Model accuracy differed significantly across both information-based (p < 0.001) and case-based questions (p = 0.029).
Accuracy peaked in Oral and Maxillofacial Surgery, Paediatric Dentistry, Periodontology, and Restorative Dentistry, but was comparatively lower in Prosthetic Dentistry, Orthodontics, and Endodontics.

Cite This Study

Kaya et al. (2026) studied this question.

synapsesocial.com/papers/6a9bd4216b95aff0620eb895 https://doi.org/10.1111/eje.70295

View Full Paper

Ask AI

Bookmark

September 5, 2026European Journal Of Dental EducationOpen Access

Evaluating the Accuracy of Large Language Models in Dentistry: A Multi‐Model Study Using Clinical Questions From Turkey's Dental Specialty Exams

View Full Paper

Ask AI

Bookmark

Authors

NKNihan Kaya FŞFatih Şengül PÇPeriş Çelikel

Discussion

Member takes

Overview

Comparative evaluation reveals variable accuracy among five large language models answering dental specialty exam questions, indicating promise as educational aids.

Key Points

To evaluate and compare the accuracy of five large language models in answering clinical questions from Turkey's dental specialty examination (DUS).
Evaluated 1,026 clinical science questions from 13 Turkish dental specialty examinations (2012–2021) across five LLMs: ChatGPT-5, Gemini 3.5 Flash, Microsoft Copilot, DeepSeek-R1, and Grok 4.
Categorized questions by dental specialty and question type (information-based vs. case-based), scoring single-prompt outputs against official answer keys using chi-square and Z tests with Bonferroni correction.
Overall accuracy ranged from 79.2% to 90.6%, with Gemini 3.5 Flash (90.6%) and ChatGPT-5 (88.2%) achieving the highest performance, while Microsoft Copilot, DeepSeek-R1, and Grok 4 were significantly less accurate (p < 0.001).
Model accuracy differed significantly across both information-based (p < 0.001) and case-based questions (p = 0.029).
Accuracy peaked in Oral and Maxillofacial Surgery, Paediatric Dentistry, Periodontology, and Restorative Dentistry, but was comparatively lower in Prosthetic Dentistry, Orthodontics, and Endodontics.

Cite This Study

Kaya et al. (2026) studied this question.

synapsesocial.com/papers/6a9bd4216b95aff0620eb895 https://doi.org/10.1111/eje.70295

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

View Full Paper

Ask AI

Bookmark