List your company

Closed-format Question Answering on Abdomen (n=143)

84.6AccuracyGPT-5.4
42.79253.64664.575.354Jul 7, 2026

Evaluation results

MethodLinks
GPT-5.42026.0784.6
HR1.52026.0780.4
Gemini-3-Flash-Preview2026.0779
HR1.5+2026.0774.8
HR12026.0774.1
Claude Opus 4.72026.0760.8
MedGemma-4B2026.0760.5
LLaMa-4-Scout2026.0758.7
MedGemma-27B2026.0744.4