Comparative performance of ChatGPT, Gemini, and Deepseek on endodontic exam questions in Turkish and English

Puntos clave

• ChatGPT, Gemini, and Deepseek show varying results on endodontic exam questions in both Turkish and English.
• Gemini performed best with a score of 85% accuracy across language pairs, while Deepseek followed at 80%.
• Assessment of AI language models focused on responses to endodontic questions, analyzing their accuracy and applicability.
• Findings highlight that the performance of AI in educational settings may differ significantly based on language.