Medical Diagnosis on MIMIC-IV (test)
78.33AccuracyGLEAN (N=3)
Evaluation Results
| Method | Links | |
|---|---|---|
| GLEAN (N=3)Agent Backbone=Qwen3-30B, Selection Strategy=BoN-162026.03 | 78.33 | |
| GLEAN (N=3)Agent Backbone=Qwen3-30B, Selection Strategy=BoN-82026.03 | 76.71 | |
| GLEAN (N=3)Agent Backbone=Qwen2.5-7B, Selection Strategy=BoN-162026.03 | 76.6 | |
| GLEAN (N=3)Agent Backbone=Qwen2.5-7B, Selection Strategy=BoN-82026.03 | 74.29 | |
| GLEAN (N=3)Agent Backbone=Qwen3-30B, Selection Strategy=BoN-42026.03 | 73.83 | |
| GLEAN (N=1)Agent Backbone=Qwen2.5-7B, Selection Strategy=BoN-162026.03 | 73.6 | |
| GLEAN (N=1)Agent Backbone=Qwen2.5-7B, Selection Strategy=BoN-82026.03 | 73.01 | |
| RAGAgent Backbone=Qwen2.5-7B, Selection Strategy=BoN-162026.03 | 72.4 | |
| GLEAN (N=1)Agent Backbone=Qwen3-30B, Selection Strategy=BoN-82026.03 | 72.37 | |
| GLEAN (N=1)Agent Backbone=Qwen3-30B, Selection Strategy=BoN-42026.03 | 71.98 | |
| GLEAN (N=1)Agent Backbone=Qwen3-30B, Selection Strategy=BoN-162026.03 | 71.67 | |
| RAGAgent Backbone=Qwen2.5-7B, Selection Strategy=BoN-82026.03 | 71.4 | |
| GLEAN (N=3)Agent Backbone=Qwen2.5-7B, Selection Strategy=BoN-42026.03 | 70.52 | |
| Self-ConsistencyAgent Backbone=Qwen3-30B, Selection Strategy=BoN-162026.03 | 70 | |
| ORMAgent Backbone=Qwen3-30B, Selection Strategy=BoN-82026.03 | 69.99 | |
| GLEAN (N=1)Agent Backbone=Qwen2.5-7B, Selection Strategy=BoN-42026.03 | 69.96 | |
| ORMAgent Backbone=Qwen3-30B, Selection Strategy=BoN-162026.03 | 69.66 | |
| Self-ConsistencyAgent Backbone=Qwen2.5-7B, Selection Strategy=BoN-162026.03 | 69.4 | |
| RAGAgent Backbone=Qwen3-30B, Selection Strategy=BoN-82026.03 | 69.33 | |
| RAGAgent Backbone=Qwen3-30B, Selection Strategy=BoN-42026.03 | 69.01 | |
| RAGAgent Backbone=Qwen3-30B, Selection Strategy=BoN-162026.03 | 69 | |
| RAGAgent Backbone=Qwen2.5-7B, Selection Strategy=BoN-42026.03 | 68.91 | |
| Self-ConsistencyAgent Backbone=Qwen3-30B, Selection Strategy=BoN-82026.03 | 68.65 | |
| ORMAgent Backbone=Qwen3-30B, Selection Strategy=BoN-42026.03 | 68.01 | |
| Self-VerificationAgent Backbone=Qwen3-30B, Selection Strategy=BoN-42026.03 | 67.43 | |
| Self-ConsistencyAgent Backbone=Qwen2.5-7B, Selection Strategy=BoN-82026.03 | 66.17 | |
| Self-VerificationAgent Backbone=Qwen3-30B, Selection Strategy=BoN-82026.03 | 66.17 | |
| MedPRMAgent Backbone=Qwen3-30B, Selection Strategy=BoN-162026.03 | 66 | |
| MedPRMAgent Backbone=Qwen3-30B, Selection Strategy=BoN-82026.03 | 65.82 | |
| Self-ConsistencyAgent Backbone=Qwen3-30B, Selection Strategy=BoN-42026.03 | 65.25 | |
| Self-VerificationAgent Backbone=Qwen2.5-7B, Selection Strategy=BoN-82026.03 | 64.74 | |
| MedPRMAgent Backbone=Qwen3-30B, Selection Strategy=BoN-42026.03 | 64.68 | |
| LLM-as-a-JudgeAgent Backbone=Qwen3-30B, Selection Strategy=BoN-162026.03 | 64.67 | |
| Self-VerificationAgent Backbone=Qwen2.5-7B, Selection Strategy=BoN-162026.03 | 64.4 | |
| ORMAgent Backbone=Qwen2.5-7B, Selection Strategy=BoN-162026.03 | 64.4 | |
| LLM-as-a-JudgeAgent Backbone=Qwen3-30B, Selection Strategy=BoN-42026.03 | 64.24 | |
| LLM-as-a-JudgeAgent Backbone=Qwen3-30B, Selection Strategy=BoN-82026.03 | 63.96 | |
| Self-VerificationAgent Backbone=Qwen2.5-7B, Selection Strategy=BoN-42026.03 | 63.66 | |
| P(True)Agent Backbone=Qwen3-30B, Selection Strategy=BoN-42026.03 | 63.22 | |
| P(True)Agent Backbone=Qwen3-30B, Selection Strategy=BoN-82026.03 | 62.45 | |
| Self-ConsistencyAgent Backbone=Qwen2.5-7B, Selection Strategy=BoN-42026.03 | 62.09 | |
| LLM-as-a-JudgeAgent Backbone=Qwen2.5-7B, Selection Strategy=BoN-162026.03 | 62 | |
| Self-VerificationAgent Backbone=Qwen3-30B, Selection Strategy=BoN-162026.03 | 62 | |
| ORMAgent Backbone=Qwen2.5-7B, Selection Strategy=BoN-82026.03 | 61.59 | |
| P(True)Agent Backbone=Qwen3-30B, Selection Strategy=BoN-162026.03 | 61.33 | |
| LLM-as-a-JudgeAgent Backbone=Qwen2.5-7B, Selection Strategy=BoN-82026.03 | 60.62 | |
| ORMAgent Backbone=Qwen2.5-7B, Selection Strategy=BoN-42026.03 | 59.65 | |
| MedPRMAgent Backbone=Qwen2.5-7B, Selection Strategy=BoN-162026.03 | 58.4 | |
| LLM-as-a-JudgeAgent Backbone=Qwen2.5-7B, Selection Strategy=BoN-42026.03 | 58.18 | |
| MedPRMAgent Backbone=Qwen2.5-7B, Selection Strategy=BoN-82026.03 | 57.77 | |
| MedPRMAgent Backbone=Qwen2.5-7B, Selection Strategy=BoN-42026.03 | 57.23 | |
| P(True)Agent Backbone=Qwen2.5-7B, Selection Strategy=BoN-82026.03 | 54.02 | |
| P(True)Agent Backbone=Qwen2.5-7B, Selection Strategy=BoN-42026.03 | 53.86 | |
| P(True)Agent Backbone=Qwen2.5-7B, Selection Strategy=BoN-162026.03 | 53.8 |