JMed48k (Text-Only) - Nurse: leaderboard
Metric: Accuracy (%) on the text-only scored items of the Japanese Nurse 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.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 2.5 Pro | 97.8 |
| 2 | GPT-5 | 96.2 |
| 3 | Claude Sonnet 4 | 95.6 |
| 4 | Gemini 2.5 Flash (Non-reasoning) | 93.6 |
| 5 | GPT-5 Mini | 92.9 |
| 6 | Gemma 4 31B (IT) | 92.5 |
| 7 | Qwen 3.5 397B A17B (Non-reasoning) | 92.3 |
| 8 | Grok 4.20 | 90.4 |
| 9 | Qwen 3.5 27B (Non-reasoning) | 89.8 |
| 10 | DeepSeek R1 | 88.2 |
| 11 | Llama 4 Maverick | 85.2 |
| 12 | Qwen 3.5 9B (Non-reasoning) | 79.8 |
| 13 | Lingshu-32B | 78.3 |
| 14 | MedGemma-27B-IT | 75.6 |
| 15 | MedGemma-4B-IT | 49.4 |
Interactive version: theaggregate.ai/benchmark?slug=jmed48k-text-only-nurse · How It Works · Data refreshed daily, snapshot 2026-10-07.