List your company

Tool-augmented Question Answering on MedQA

74.11AccuracyEPC-AW
66.8368.7270.6172.5May 22, 2026

Evaluation results

MethodLinks
EPC-AWBackbone=Qwen3-Coder-30B2026.0574.11—
RollbackBackbone=Qwen3-Coder-30B2026.0571.562.55
RetryBackbone=Qwen3-Coder-30B2026.05713.11
No-Repair (AgentFlow)Backbone=Qwen3-Coder-30B2026.0567.117