List your company

Question Answering on MMLU-CK (test)

60.95AccuracyPubMed Reasoner
51.038853.611956.18558.7581Mar 28, 2026

Evaluation results

MethodLinks
PubMed ReasonerBackbone=Qwen2.5-72B2026.0360.95
Self-reflection agentBackbone=Qwen2.5-72B2026.0359.67
Without retrievalBackbone=Qwen2.5-72B2026.0358.36
Without retrievalBackbone=Qwen2.5-7B2026.0356.64
Self-reflection agentBackbone=Qwen2.5-7B2026.0356.11
PubMed ReasonerBackbone=Qwen2.5-7B2026.0356.09
Without retrievalBackbone=Qwen2.5-1.5B2026.0355.67
Self-reflection agentBackbone=Qwen2.5-1.5B2026.0351.44
PubMed ReasonerBackbone=Qwen2.5-1.5B2026.0351.42