medbench.kr

Korean clinical knowledge (KMLE-level MCQ) preview

clinicalqa@0 · mcq · primary metric: accuracy · spec hash 77e37cd0ca6b

ClinicalQA has been public on Hugging Face since 2025-04-04. Models trained after that date may have seen it.

#ModelTypeTieraccuracy (95% CI)nCost (USD)Date
1 GLM-5.3 753B-A40B
kgpu/glm-5.3-ax-753b-a40b
llm baseline 0.909 (0.892–0.926) 1045 0.0000 2026-10-29
2 Gemini 3.8 Flash
openrouter/google/gemini-3.8-flash
llm baseline 0.908 (0.891–0.925) 1045 2.3561 2026-10-29
3 DeepSeek V4.1 Flash (2026-10-17)
openrouter/deepseek/deepseek-v4.1-flash-1017
llm baseline 0.904 (0.887–0.922) 1045 0.4731 2026-10-29
4 Qwen3.6 35B-A3B
kgpu/qwen3.6-35b-a3b
llm baseline 0.903 (0.885–0.922) 1045 2.2815 2026-10-29
5 DeepSeek V4.1 Flash 552B-A16B
kgpu/deepseek-v4.1-flash-ax-552b-a16b
llm baseline 0.900 (0.883–0.918) 1045 0.0000 2026-10-29
6 GPT-6 Luna
openrouter/openai/gpt-6-luna
llm baseline 0.900 (0.882–0.919) 1045 0.1103 2026-10-29
7 DeepSeek V4 Pro 1.6T-A49B
kgpu/deepseek-v4-pro-ax-1.6t-a49b
llm baseline 0.894 (0.876–0.911) 1045 0.0000 2026-10-29
8 Qwen3.8 Flash Next 180B-A6B
kgpu/qwen3.8-flash-next-180b-a6b
llm baseline 0.894 (0.877–0.913) 1045 0.0000 2026-10-29
9 Claude Haiku 5.5
openrouter/anthropic/claude-haiku-5.5
llm baseline 0.885 (0.866–0.903) 1045 0.4409 2026-10-29
10 Gemma 4 26B-A4B
kgpu/gemma-4-26b-a4b
llm baseline 0.874 (0.854–0.894) 1045 0.2035 2026-10-29
11 MedGemma 1.5 4B
kgpu/medgemma-1.5-4b
llm baseline 0.765 (0.739–0.789) 1045 0.1244 2026-10-29
12 Majority class (always B)
medbench/majority-b
baseline baseline 0.343 (0.314–0.370) 1045 0.0000 2026-10-29

Data

ClinicalQA: Korean Medical Knowledge QA Benchmark (text). License: CC BY 4.0 (upstream Apache-2.0). Source: https://doi.org/10.23421/2026.0800

SNUH-HARI. ClinicalQA: Korean Medical Knowledge QA Benchmark. Seoul National University Hospital Healthcare AI Research Institute, 2025. doi:10.23421/2026.0800

Task specification

{
  "dataset": "clinicalqa",
  "kind": "mcq",
  "preview": true,
  "primary_metric": "accuracy",
  "prompt": "clinicalqa-p0",
  "slug": "clinicalqa",
  "source": {
    "file": "data/train-00000-of-00001.parquet",
    "repo": "snuh/ClinicalQA",
    "revision": "29bdc9043f3e693b7671f868e2038ec7ac288acd"
  },
  "split": "test-only (all 1,045 items)",
  "title": "Korean clinical knowledge (KMLE-level MCQ)",
  "version": 0
}

JSON: /v1/tasks/clinicalqa · /v1/leaderboard?task=clinicalqa@0