List your company

Radiology Report Meta-Evaluation on ReEvalMed

100D ScoreHulu-Med-32B
15.7637.6359.581.37Jun 17, 2026

Evaluation results

MethodLinks
Hulu-Med-32BType=Open-source, Passes=1-Pass2026.061002361.577
Hulu-Med-32BType=Open-source, Inference Strategy=1-Pass2026.061002361.577
LingShu-32BType=Open-source, Passes=1-Pass2026.0699.56481.835.5
LingShu-32BType=Open-source, Inference Strategy=1-Pass2026.0699.56481.835.5
GPT-5.1Type=Proprietary, Passes=1-Pass2026.0698.542.570.556
GPT-5.1Type=Proprietary, Inference Strategy=1-Pass2026.0698.542.570.556
Claude Sonnet 4.5Type=Proprietary, Passes=2-Pass2026.06985174.547
Claude Sonnet 4.5Type=Proprietary, Inference Strategy=2-Pass2026.06985174.547
LingShu-32BType=Open-source, Passes=2-Pass2026.0697.55.551.592
LingShu-32BType=Open-source, Inference Strategy=2-Pass2026.0697.55.551.592
Qwen3-MaxType=Open-source, Passes=1-Pass2026.069753.575.343.5
Qwen3-MaxType=Open-source, Inference Strategy=1-Pass2026.069753.575.343.5
Qwen3-MaxType=Open-source, Passes=2-Pass2026.0696.54570.851.5
Qwen3-8BType=Open-source, Passes=1-Pass2026.0696.5852.388.5
Qwen3-8BType=Open-source, Inference Strategy=1-Pass2026.0696.5852.388.5
Qwen3-MaxType=Open-source, Inference Strategy=2-Pass2026.0696.54570.851.5
Hulu-Med-32BType=Open-source, Passes=2-Pass2026.069652.574.343.5
Hulu-Med-32BType=Open-source, Inference Strategy=2-Pass2026.069652.574.343.5
Qwen3-8BType=Open-source, Passes=2-Pass2026.0695.5952.386.5
Qwen3-8BType=Open-source, Inference Strategy=2-Pass2026.0695.5952.386.5
GPT-5.1Type=Proprietary, Passes=2-Pass2026.069569.582.325.5
GPT-5.1Type=Proprietary, Inference Strategy=2-Pass2026.069569.582.325.5
Gemini 3 ProType=Proprietary, Passes=1-Pass2026.069457.575.836.5
Gemini 3 ProType=Proprietary, Inference Strategy=1-Pass2026.069457.575.836.5
Claude Sonnet 4.5Type=Proprietary, Passes=1-Pass2026.0691.58487.87.5
Claude Sonnet 4.5Type=Proprietary, Inference Strategy=1-Pass2026.0691.58487.87.5
FineRadScoreType=LLM, Threshold=Maximin, Backbone=GPT-5.12026.0686.568.577.518
RadFactType=LLM, Threshold=Maximin, Backbone=GPT-5.12026.0680566824
CRIMSONType=LLM, Threshold=Maximin2026.067580.577.85.5
MedGemma-4BType=Open-source, Passes=2-Pass2026.066767.567.30.5
MedGemma-4BType=Open-source, Inference Strategy=2-Pass2026.066767.567.30.5
Gemini 3 ProType=Proprietary, Passes=2-Pass2026.066579.572.314.5
Gemini 3 ProType=Proprietary, Inference Strategy=2-Pass2026.066579.572.314.5
GREENType=LLM, Threshold=Maximin2026.0653.541.547.512
AlignScoreType=NLP, Threshold=Maximin2026.065151510
MedGemma-4BType=Open-source, Passes=1-Pass2026.064992940
MedGemma-4BType=Open-source, Inference Strategy=1-Pass2026.064992940
CheXbertType=Med., Threshold=Maximin2026.064749482
RaTEScoreType=Med., Threshold=Maximin2026.0635.535.535.50
RadGraphType=Med., Threshold=Maximin2026.0628.528.528.50
RadBERTScoreType=Med., Threshold=Maximin2026.0626.526.526.50
BERTScoreType=NLP, Threshold=Maximin2026.062424240
BLEUType=NLP, Threshold=Maximin2026.061919190