LLMs perform worse on dental licensing examinations than on medical licensing examinations¶

Claim · All claims
Evidence · 1 study · quant-synthesis r2 · q3
Currency · newest study 2025 · not yet checked against its sources · not reviewed by a person · recheck by 2028: tool and model accuracy changes quickly

Subclaims¶

q3 i? The 54% accuracy of GPT-3.5 and 72% of GPT-4 in dental licensing examinations are both lower than the accuracy rates reported in a meta-analysis on medical licensing examinations (GPT-3.5: 59%, GPT-4: 81%). → Liu Mingxin 2025

Evidence¶

Liu Mingxin 2025¶

Liu Mingxin, Okuhara Tsuyoshi, Huang Wenbo, Ogihara Atsushi, Nagao Hikari Sophia, Okada Hiroko, Kiuchi Takahiro. (2025). Large Language Models in Dental Licensing Examinations: Systematic Review and Meta-Analysis. International Dental Journal. https://doi.org/10.1016/j.identj.2024.10.014

q3 · i? · quant-synthesis · r2

Comparison reported in the discussion section of this systematic review, contrasting its dental examination results with a prior meta-analysis on medical licensing examinations. The review reports "54% accuracy rate of GPT-3 and 72% of GPT-4 in dental licensing examinations are both lower than the accuracy rates reported in another meta-analysis on LLMs' performance in medical licensing examinations".

"54% accuracy rate of GPT-3 and 72% of GPT-4 in dental licensing examinations are both lower than the accuracy rates reported in another meta-analysis on LLMs' performance in medical licensing examinations (GPT-3.5: 59%, GPT-4: 81%)."

Discussion¶

Related Claims¶

• GPT-3.5 achieves an integrated accuracy of 54% on dental licensing examinations — related
• GPT-3.5 performs significantly better on dental licensing examinations from English-speaking countries than non-English-speaking countries — related
• LLMs provide more detailed explanations for incorrect answers than correct ones in dental licensing examinations — related
• Subgroup analyses show significant between-model differences, with Copilot (75.25%) and ChatGPT-4 (72.97%) achieving the highest pooled accuracies — related
• GPT-4 outperforms GPT-3.5 and Bard on dental licensing examinations, achieving an integrated accuracy of 72% — related
• Large language models achieve a pooled accuracy of 72% on medical licensing examination questions across 36 studies — related
• Studies using both text- and image-based questions show higher pooled LLM accuracy (78%) than text-only studies (71%) — related
• GPT-4 achieves passing accuracy on USMLE Steps 1–3 (mean 86%), significantly exceeding GPT-3.5 (mean 56%) — related
• Newer OpenAI LLM versions achieve significantly higher pooled accuracy on biomedical MCQs than earlier versions (GPT‐4o–o1: 90%; GPT‐4: 82%; GPT‐3/3.5: 59%) — related
• ChatGPT-4.0 outperforms ChatGPT-3.5 in medical licensing exams, achieving a pooled accuracy of 81.8% versus 60.8% — related