Medical Question Answering on MedQA In-Distribution
92.19AccuracySkillFlow
Evaluation Results
| Method | Links | |
|---|---|---|
| SkillFlowMethod Category=Ours2026.05 | 92.19 | |
| FlowSteerMethod Category=Agent+RL2026.05 | 90.63 | |
| Qwen3.5-AFlowMethod Category=AFlow, Backbone=Qwen3.52026.05 | 89.84 | |
| AgentFlowMethod Category=Agent+RL2026.05 | 87.5 | |
| SkillRLMethod Category=Agent+RL2026.05 | 85.94 | |
| v4-flashMethod Category=Baseline2026.05 | 83.59 | |
| Qwen3.5-GRPOMethod Category=GRPO, Backbone=Qwen3.52026.05 | 77.34 | |
| Qwen3.5-SFTMethod Category=SFT, Backbone=Qwen3.52026.05 | 73.44 | |
| Qwen3.5-9BMethod Category=Baseline2026.05 | 69.53 |