TCM clinical reasoning and prescription evaluation on 1,000 clinical cases (val)
42.96Total ScoreCORE-Acu
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| CORE-Acuprotocol=LLM-as-a-judge (GPT-5.2)2026.03 | 42.96 | 8.93 | 8.28 | 8.74 | 8.16 | 8.85 | |
| GLM-4.6protocol=LLM-as-a-judge (GPT-5.2)2026.03 | 42.26 | 8.77 | 8.34 | 8.51 | 9.04 | 7.6 | |
| Qwen-Maxprotocol=LLM-as-a-judge (GPT-5.2)2026.03 | 36.88 | 7.52 | 7.16 | 6.99 | 7.3 | 7.91 | |
| GPT-4oprotocol=LLM-as-a-judge (GPT-5.2)2026.03 | 36.73 | 7.54 | 7.2 | 7.09 | 7.24 | 7.66 | |
| Qwen3-8Bprotocol=LLM-as-a-judge (GPT-5.2)2026.03 | 29.06 | 5.93 | 5.75 | 5.62 | 5.34 | 6.43 | |
| HuatuoGPTprotocol=LLM-as-a-judge (GPT-5.2)2026.03 | 20 | 3.99 | 3.92 | 3.77 | 3.64 | 4.68 |