List your company

Radiology Report Evaluation on Open-i 50-sample benchmark

100D ScoreRadFact
91.6893.849698.16Jun 17, 2026

Evaluation results

MethodLinks
RadFactType=Baseline2026.061001001000
FineRadScoreType=Baseline2026.061001001000
Claude Sonnet 4.5Type=Proprietary2026.061001001000
Qwen3-8BType=Open-source2026.06100768824
Qwen3-8B + SFT + RLType=Open-source2026.0610096984
MedGemma-4BType=Open-source2026.0610092968
MedGemma-4B + SFT + RLType=Open-source2026.061001001000
CRIMSONType=Baseline2026.0692100968