{"rank":null,"reference":false,"run_id":"7993fc36d3a5","task":"clinicalqa@0","model":{"id":"medgemma-1.5-4b","slug":"kgpu/medgemma-1.5-4b","name":"MedGemma 1.5 4B","type":"llm","developer":"Google"},"tier":"baseline","metric":{"name":"accuracy","value":0.7645933014354067,"ci_low":0.738755980861244,"ci_high":0.7894736842105263},"n":1045,"parse_rate":0.9952153110047847,"cost_usd":0.12438564,"eval_date":"2026-10-29","url":"https://medbench.kr/r/7993fc36d3a5","status":"done","listed":true,"current_spec":true,"metrics":{"accuracy":{"value":0.7645933014354067,"ci_low":0.738755980861244,"ci_high":0.7894736842105263,"n":1045,"n_unparsed":5},"parse_rate":0.9952153110047847,"error_rate":0.0,"by_category":{"Allergy and Immunology":{"value":0.7222222222222222,"n":36},"Cardiology":{"value":0.8539325842696629,"n":89},"Dermatology":{"value":0.5833333333333334,"n":12},"Emergency Medicine":{"value":0.78125,"n":32},"Endocrinology":{"value":0.8235294117647058,"n":68},"Endocrinology, metabolic, and nutritional disorders":{"value":1.0,"n":1},"Forensic Medicine":{"value":1.0,"n":1},"Gastroenterology":{"value":0.8214285714285714,"n":112},"General Surgery":{"value":0.6896551724137931,"n":58},"Geriatric Medicine":{"value":1.0,"n":1},"Gynecology":{"value":0.5857142857142857,"n":70},"Hematology and Oncology":{"value":0.8958333333333334,"n":48},"Infectious Diseases":{"value":0.6923076923076923,"n":52},"Musculoskeletal (e.g., costochondritis)":{"value":1.0,"n":1},"Nephrology":{"value":0.7207207207207207,"n":111},"Neurology / Neurosurgery":{"value":0.7959183673469388,"n":49},"Obstetrics":{"value":0.8333333333333334,"n":18},"Ophthalmology":{"value":1.0,"n":1},"Orthopedic Surgery":{"value":0.6666666666666666,"n":24},"Otolaryngology":{"value":0.8571428571428571,"n":14},"Otolaryngology (ENT)":{"value":0.7857142857142857,"n":14},"Pediatrics":{"value":0.7321428571428571,"n":56},"Pharmacology":{"value":1.0,"n":1},"Psychiatry":{"value":0.6888888888888889,"n":45},"Pulmonology (Respiratory Medicine)":{"value":0.8307692307692308,"n":65},"Radiology":{"value":1.0,"n":1},"Rheumatology":{"value":0.7755102040816326,"n":49},"Urology":{"value":0.7692307692307693,"n":13},"Vascular Surgery":{"value":1.0,"n":2},"nan":{"value":1.0,"n":1}}},"config":{"max_tokens":16384,"prompt":"clinicalqa-p0","temperature":0},"run_key":"dca8c80bee5d4f700d225c14e5d8cdd2c15dbc4ecc3d801025230c74f4a9b6b1","harness_version":"p0.1","task_spec_hash":"77e37cd0ca6bded7fba03792ec1922342b3f6a7dc3ce602881ce196e6aa7377c","model_id":"medgemma-1.5-4b","tokens_in":399078,"tokens_out":1124133,"dataset":{"slug":"clinicalqa","title":"ClinicalQA: Korean Medical Knowledge QA Benchmark","license":"CC BY 4.0 (upstream Apache-2.0)","license_url":"https://creativecommons.org/licenses/by/4.0/","citation":"SNUH-HARI. ClinicalQA: Korean Medical Knowledge QA Benchmark. Seoul National University Hospital Healthcare AI Research Institute, 2025. doi:10.23421/2026.0800"},"predictions_url":"https://medbench.kr/artifacts/7993fc36d3a5/predictions.parquet","bibtex":"@misc{medbench_7993fc36d3a5,\n  title = {MedGemma 1.5 4B on Korean clinical knowledge (KMLE-level MCQ) (clinicalqa@0)},\n  author = {{medbench.kr}},\n  year = {2026},\n  howpublished = {\\url{https://medbench.kr/r/7993fc36d3a5}},\n  note = {Evaluated 2026-10-29, harness p0.1, run key dca8c80bee5d}\n}"}