Comparative performance of GPT-4, GPT-o3, GPT-5, Gemini-3-Flash, and DeepSeek-R1 in ophthalmology question answering
GPT-o3 and Gemini-3-Flash achieve superior stability and accuracy in ophthalmology Question Answering (QA), making them suitable for high-stakes clinical decision support. The open-source model DeepSeek-R1 shows competitive potential, especially in complex tasks. Notably, GPT-5 failed to surpass its predecessor in both accuracy and consistency in this specialized domain. Prompt engineering has a limited impact on performance for closed-ended medical questions. Future work should extend to multimodal integration and real-world clinical validation to enhance the practical utility and reliability of LLMs in medicine.
Authors
- Weihua Yang (ORCID: https://orcid.org/0000-0002-7629-0193)
- Xiaoqin Wang (ORCID: https://orcid.org/0000-0003-4404-0082)
- Wei Chi
- Jiaoman Wang
- Xinya Hu
- Ping Zhang
- Xianming Fan
Institutions
- Wenzhou Medical University (CN)
- Affiliated Eye Hospital of Wenzhou Medical College (CN)
- Second Hospital of Shandong University (CN)
- Southern Medical University Shenzhen Hospital (CN)
- He Eye Hospital (CN)
- Shenzhen Second People's Hospital (CN)
Publication Details
- Journal
- Frontiers in Cell and Developmental Biology
- Published
- 2026-01-29
- DOI
- https://doi.org/10.3389/fcell.2026.1744389
- Primary Topic
- Artificial Intelligence in Healthcare and Education
- Type
- article
- Field-Weighted Citation Impact
- 0.00
Funders
- National Natural Science Foundation of China
- Sanming Project of Medicine in Shenzhen
- National Key Research and Development Program of China