Clinical Medical Examination Question Answering on CMExam (val)
77.65AccuracySafeMed-R1
Evaluation Results
| Method | Links | |
|---|---|---|
| SafeMed-R1Optimization=CoT reasoning, safety-ethics alignment, improved GRPO2026.05 | 77.65 | |
| Qwen3-32B+SFT2026.05 | 76.11 | |
| qwen3-30B-A3B-Instruct-25072026.05 | 75.64 | |
| Qwen3-32B2026.05 | 75.39 | |
| Baichuan-M2-32B2026.05 | 75.36 | |
| gemini-2.5-flash2026.05 | 71.27 | |
| gpt-5-mini2026.05 | 68.65 | |
| DeepSeek-R1-Distill-Qwen-32B2026.05 | 67.04 | |
| Qwen2.5-32B-Instruct2026.05 | 66.5 | |
| HuatuoGPT-o1-70B2026.05 | 54.1 | |
| google_medgemma-27b-text-it2026.05 | 53.43 | |
| Lingshu-32B2026.05 | 50.36 |