JMed48k (Text-Only) - Midwife: leaderboard
Metric: Accuracy (%) on the text-only scored items of the Japanese Midwife 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 | 94.3 |
| 2 | GPT-5 | 90.4 |
| 3 | Claude Sonnet 4 | 86.8 |
| 4 | GPT-5 Mini | 84.4 |
| 5 | Gemini 2.5 Flash (Non-reasoning) | 84.4 |
| 6 | Qwen 3.5 397B A17B (Non-reasoning) | 81.1 |
| 7 | Gemma 4 31B (IT) | 80.3 |
| 8 | DeepSeek R1 | 79.5 |
| 9 | Qwen 3.5 27B (Non-reasoning) | 78.7 |
| 10 | Grok 4.20 | 78.3 |
| 11 | Llama 4 Maverick | 72.4 |
| 12 | Qwen 3.5 9B (Non-reasoning) | 66.5 |
| 13 | MedGemma-27B-IT | 61.8 |
| 14 | Lingshu-32B | 59.6 |
| 15 | MedGemma-4B-IT | 35.6 |
Interactive version: theaggregate.ai/benchmark?slug=jmed48k-text-only-midwife · How It Works · Data refreshed daily, snapshot 2026-10-07.