Medical Diagnosis on MedEinst Robust 1.0
24.21Robust AccuracyECR-Agent (Qwen3-32B)
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| ECR-Agent (Qwen3-32B)Size=32B2026.01 | 24.21 | 33.75 | |
| GPT-5Size=-2026.01 | 15.78 | 51.87 | |
| Kimi-k2Size=-2026.01 | 12.46 | 47.12 | |
| Claude-Sonnet-4.5Size=-2026.01 | 12.36 | 42.98 | |
| Qwen/Qwen3-32BSize=32B2026.01 | 11.86 | 43.46 | |
| MedGemma-27B-text-itSize=27B2026.01 | 11.68 | 53.88 | |
| Qwen/Qwen3-235B-A22BSize=235B2026.01 | 11.4 | 43.32 | |
| DeepSeek-R1Size=-2026.01 | 11.32 | 46.12 | |
| Qwen/Qwen3-14BSize=14B2026.01 | 11.28 | 54.19 | |
| ZhipuAI/GLM-4.6Size=-2026.01 | 11.25 | 47.63 | |
| Qwen/QwQ-32BSize=32B2026.01 | 11.14 | 44.88 | |
| Gemini-2.5-proSize=-2026.01 | 10.97 | 60.9 | |
| MDAgent (Qwen3-32B)Size=32B2026.01 | 10.34 | 40.34 | |
| DyLAN (Qwen3-32B)Size=32B2026.01 | 8.11 | 41.69 | |
| Baichuan-M2-32BSize=32B2026.01 | 7.18 | 66.1 | |
| Lingshu-32BSize=32B2026.01 | 6.17 | 54.2 | |
| Lingshu-7BSize=7B2026.01 | 3.2 | 36.74 | |
| Llama3-Med42-8BSize=8B2026.01 | 1.19 | 44 |