IgakuQA119: leaderboard
Japanese medical licensing exam benchmark with 500 questions from the 119th Japan Medical Licensing Examination, testing medical knowledge in Japanese.
Metric: Accuracy (%). Source: github.com. Saturation forecast: Estimated already saturated. 27 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.5 Pro | 97 | #152 |
| 2 | O3 | 96.4 | #124 |
| 3 | Gemini 2.5 Flash | 95.6 | #245 |
| 4 | Claude Sonnet 4 | 94.2 | #199 |
| 5 | Qwen 3 235B A22B | 92.4 | #313 |
| 6 | DeepSeek R1 0528 | 92.2 | #222 |
| 7 | DeepSeek R1 | 89.6 | #251 |
| 8 | Llama 4 Maverick | 88 | #464 |
| 9 | Gemini 2.0 Flash | 87.2 | #342 |
| 10 | QwQ-32B | 86 | #421 |
| 11 | Qwen 3 32B | 83 | #443 |
| 12 | Qwen 3 30B A3B | 82.4 | #493 |
| 13 | DeepSeek V3 (0324) | 79.8 | #340 |
| 14 | Qwen 2.5 72B | 79.6 | #499 |
| 15 | Llama 4 Scout | 78.4 | #656 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=igakuqa119 · How It Works · Data refreshed daily, snapshot 2026-10-20.