Capability of GPT-4V(ision) in the Japanese National Medical Licensing Examination: Evaluation Study
Open accessThe DOI matches its Crossref record; 7 of the 7 references it cites were found in a registry or an index.
Abstract
Background: Previous research applying large language models (LLMs) to medicine was focused on text-based information. Recently, multimodal variants of LLMs acquired the capability of recognizing images. Objective: We aim to evaluate the image recognition capability of generative pretrained transformer (GPT)-4V, a recent multimodal LLM developed by OpenAI, in the medical field by testing how visual information affects its performance to answer questions in the 117th Japanese National Medical Licensing Examination. Methods: We focused on 108 questions that had 1 or more images as part of a question and presented GPT-4V with the same questions under two conditions: (1) with both the question text and associated images and (2) with the question text only. We then compared the difference in accuracy between the 2 conditions using the exact McNemar test. Results: Among the 108 questions with images, GPT-4V’s accuracy was 68% (73/108) when presented with images and 72% (78/108) when presented without images (P=.36). For the 2 question categories, clinical and general, the accuracies with and those without images were 71% (70/98) versus 78% (76/98; P=.21) and 30% (3/10) versus 20% (2/10; P≥.99), respectively. Conclusions: The additional information from the images did not significantly improve the performance of GPT-4V in the Japanese National Medical Licensing Examination.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Language Models are Few-Shot Learners
Brown, T. B., Mann, B., Ryder, N., Subbiah, M., Kaplan, J., Dhariwal, P., Neelakantan, A., Shyam, P., Sastry, G., Askell, A., et al. (2020) · arXiv · doi:10.48550/arxiv.2005.14165 · Cited by 23,372
Sallam, M. (2023) · Healthcare · doi:10.3390/healthcare11060887 · Cited by 3,126 · cited 2×
“ChatGPT may serve as a valuable teaching assistant in medical education; however, the inaccuracies in its responses are a significant concern [5,7].”
BackgroundMIMIC-CXR, a de-identified publicly available database of chest radiographs with free-text reports
Johnson, A. E. W., Pollard, T. J., Berkowitz, S. J., Greenbaum, N. R., Lungren, M. P., Deng, C.-Y., Mark, R. G., & Horng, S. (2019) · Scientific Data · doi:10.1038/s41597-019-0322-0 · Cited by 1,816 · cited 1×
“These models use publicly available data sets that combine medical images and text, including MIMIC-CXR [21], which contains chest x-ray images with their associated reports, and PMC-OA [22], a compilation of the figures and captions from open-access medical journal papers.”
MethodVision-Language Models for Vision Tasks: A Survey
Zhang, J., Huang, J., Jin, S., & Lu, S. (2024) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2024.3369699 · Cited by 1,051
The rise of ChatGPT : Exploring its potential in medical education
Lee, H. (2024) · Anatomical Sciences Education · doi:10.1002/ase.2270 · Cited by 596 · cited 2×
“ChatGPT may serve as a valuable teaching assistant in medical education; however, the inaccuracies in its responses are a significant concern [5,7].”
BackgroundPerformance of GPT-3.5 and GPT-4 on the Japanese Medical Licensing Examination: Comparison Study
Takagi, S., Watari, T., Erabi, A., & Sakaguchi, K. (2023) · JMIR Medical Education · doi:10.2196/48002 · Cited by 271
Wu, C., Lei, J., Zheng, Q., Zhao, W., Lin, W., Zhang, X., Zhou, X., Zhao, Z., Zhang, Y., Wang, Y., et al. (2023) · arXiv · doi:10.48550/arxiv.2310.09909 · Cited by 43 · cited 2×
“According to the case studies by Wu et al [18], GPT-4V could recognize the modality and anatomy of medical images, but it could hardly make accurate diagnoses and its prediction relied heavily on the patient’s medical history.”
MethodResult
Cited by
Ullah, E., Parwani, A., Baig, M. M., & Singh, R. (2024) · Diagnostic Pathology · doi:10.1186/s13000-024-01464-7 · Cited by 281
Liu, M., Okuhara, T., Dai, Z., Huang, W., Gu, L., Okada, H., Furukawa, E., & Kiuchi, T. (2025) · International Journal of Medical Informatics · doi:10.1016/j.ijmedinf.2024.105673 · Cited by 73
Jaleel, A., Aziz, U., Farid, G., Zahid Bashir, M., Mirza, T. R., Khizar Abbas, S. M., Aslam, S., & Sikander, R. M. H. (2025) · JMIR Medical Education · doi:10.2196/68070 · Cited by 25
“) [60] Instead of providing strictly accurate information, they generate responses based on the probability of the most appropriate words given the data they have been trained on.”
Scherr, R., Spina, A., Dao, A., Andalib, S., Halaseh, F. F., Blair, S., Wiechmann, W., & Rivera, R. (2025) · JMIR Formative Research · doi:10.2196/66478 · Cited by 10
“However, acute asthma exacerbation is not an exceed-ingly complex problem and has well-established management guidelines that likely exist in ChatGPT’s training data, so its medical accuracy cannot be generalized to more complex clinical simulations like intensive care management [2,4].”
Metadata
- Authors
- Nakao, T., Miki, S., Nakamura, Y., Kikuchi, T., Nomura, Y., Hanaoka, S., Yoshikawa, T., & Abe, O.
- Venue
- JMIR Medical Education
- Publisher
- JMIR Publications Inc.
- Published
- 2024-03-12
- Volume
- 10
- Pages
- e54393
- ISSN
- 2369-3762 (ISSN-L)
- OpenAlex
- W4392702964
- Cited by
- 55
- References
- 7
- FWCI
- 4.41
- Type
- journal article
- Language
- en
- Open access
- gold
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: Capability of GPT-4V(ision) in the Japanese National Medical Licensing Examination: Evaluation Study. https://app.cochat.ai/papers/10.2196/54393
- The paper
- Nakao, T., Miki, S., Nakamura, Y., Kikuchi, T., Nomura, Y., Hanaoka, S., Yoshikawa, T., & Abe, O. (2024). Capability of GPT-4V(ision) in the Japanese National Medical Licensing Examination: Evaluation Study. JMIR Medical Education. https://doi.org/10.2196/54393
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.