ClinicalMC (Chinese): leaderboard
Metric: Mean of 11 task scores (%; triage accuracy, examination recall, diagnosis F1, LLM-judged diagnosis basis and differential diagnosis, treatment and per-course IoU; GPT-4o-mini patient and examiner agents, gold inputs from earlier stages, mean of three runs; 1,275 Chinese records from MedEureka). Source: arxiv.org. Saturation forecast: Around January 2028. 22 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 3 Next 80B A3B Instruct | 48.17 |
| 2 | GPT-5 Mini | 42.04 |
| 3 | Qwen Turbo | 38.62 |
| 4 | Qwen 2.5 32B Instruct | 38.36 |
| 5 | Qwen 2.5 7B Instruct | 38.27 |
| 6 | Llama 3.3 70B Instruct | 37.93 |
| 7 | Qwen 2.5 14B Instruct | 37.81 |
| 8 | Qwen 2.5 72B Instruct | 37.59 |
| 9 | GPT-4o Mini | 34.84 |
| 10 | Mixtral 8x22B Instruct (v0.1) | 30.25 |
| 11 | Llama 3.2 3B Instruct | 22.01 |
| 12 | Mistral 7B Instruct (v0.3) | 20.65 |
Interactive version: theaggregate.ai/benchmark?slug=clinicalmc-chinese · How It Works · Data refreshed daily, snapshot 2026-09-26.