Medical Question Answering on MedQA CN (Accuracy)
85.49AccuracySafeMed-R1
Evaluation results
| Method | Links | |
|---|---|---|
| SafeMed-R1Optimization=CoT reasoning, safety-ethics alignment, improved GRPO2026.05 | 85.49 | |
| Qwen3-32B+SFT2026.05 | 84.5 | |
| qwen3-30B-A3B-Instruct-25072026.05 | 82.95 | |
| Qwen3-32B2026.05 | 82.52 | |
| Baichuan-M2-32B2026.05 | 81.58 | |
| google_medgemma-27b-text-it2026.05 | 79.05 | |
| gemini-2.5-flash2026.05 | 78.34 | |
| DeepSeek-R1-Distill-Qwen-32B2026.05 | 76.36 | |
| gpt-5-mini2026.05 | 75.86 | |
| Qwen2.5-32B-Instruct2026.05 | 74.58 | |
| HuatuoGPT-o1-70B2026.05 | 69.12 | |
| Lingshu-32B2026.05 | 60.48 |