JMed48k (Text-Only) - Physician: leaderboard

Accuracy (%) on the text-only questions of the Japanese Physician national licensing examination, 2021-2025, in JMed48k (2026), scored against the official answers of the Ministry of Health, Labour and Welfare; higher is better.

Metric: Accuracy (%) on the text-only scored items of the Japanese Physician national licensing examination (2021-2025) in JMed48k-Eval, official MHLW gold answers, single-shot Japanese prompts at temperature 0 with reasoning disabled where the model has a switch, JSON-constrained answers; higher is better. Source: arxiv.org. Saturation forecast: Estimated already saturated. 21 models tracked. Updated 14 Nov 2026.

Top models

#ModelSettingScoreOverall rank
1Gemini 2.5 Pro98#276
2GPT-597.9#178
3GPT-5 Mini95.7#364
4Claude Sonnet 495.4#410
5Qwen 3.5 397B A17BNon-reasoning94.4#264
6Gemma 4 31B (IT)94.1#350
7Gemini 2.5 FlashNon-reasoning94#463
8DeepSeek R190.5#482
9Qwen 3.5 27BNon-reasoning90.4#375
10Grok 4.2090.1#301
11GLM-4.6V-FlashNon-reasoning87.6#1040
12Llama 4 Maverick86.5#1110
13Qwen 3.5 9BNon-reasoning79.7#828
14Lingshu-32B77#1113
15MedGemma-27B-IT72.1#1350

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=jmed48k-text-only-physician · How It Works · Data refreshed daily, snapshot 2026-11-21.