ResearchBenchmarksClinical Case Generation on Held-out (test)Follow18.98BLEU-4CORE-Acu-0.14564.81979.78514.7503Mar 9, 2026Evaluation ResultsMethodMethodLinksBLEU-4ROUGE-LEntity-F1CORE-Acu2026.0318.9833.3446.12Qwen-Maxvariant=Generalvariant=General2026.038.6519.8830.87Qwen3-8Bvariant=Basevariant=Base2026.038.1418.8330.62GPT-4o2026.037.2219.8529.06GLM-4.62026.036.9719.735.11HuatuoGPT-o1-7B2026.030.5912.0921.61