Medical Question Answering on HealthBench Hard

34.7Score

Baichuan-M2-32B

Updated 3d ago

Evaluation Results

Method
Baichuan-M2-32B 2026.02	34.7	-
Qwen3-30B-A3B-Instruct + More Query Rubrics 2026.02	33.4	-
o3 2026.02	31.6	-
Qwen3-4B-Instruct + More Query Rubrics 2026.02	27.2	-
Qwen3-4B-Instruct + Principle Rubrics 2026.02	24.4	-
Qwen3-4B-Instruct + Doctor Rubrics 2026.02	22.9	-
Qwen3-4B-Instruct + Draft Rubrics 2026.02	21.2	-
Gemini-2.5-Pro 2026.02	18.5	-
HuatuoGPT-o1-72B 2026.02	18.2	-
Qwen3-235B-Instruct 2026.02	16.2	-
GPT-4.1 2026.02	16	-
Deepseek-R1 2026.02	15.1	-
Claude-3.7-Sonnet 2026.02	15	-
Qwen3-30B-A3B-Instruct 2026.02	15	-
Qwen3-32B 2026.02	12	-
Qwen3-4B-Instruct 2026.02	5.2	-
GPT-5.2 Thinking 2026.02	-	42.04
Gemini 3.0 Pro 2026.02	-	17.7
Gemini 2.5 Pro 2026.02	-	16.8
MedXIAOHE 2026.02	-	46.1