List your company

Multiple Choice Question Answering on MMLU ProX Health

76.56Accuracy (ZH)MED-COREASONER
50.071256.948163.82570.7019Jan 13, 2026

Evaluation results

MethodLinks
MED-COREASONERModel Category=Closed-Source, Backbone=GPT-5.12026.0176.5677.8776.8577.7277.4378.678.3175.9877.42
GPT-5.2Model Category=Closed-Source2026.0172.6373.9473.9476.427775.6975.9870.674.53
MED-COREASONERModel Category=Open-Source, Backbone=DeepSeek-3.22026.0171.9171.9172.3473.3674.5273.2273.9168.8572.5
MED-COREASONERModel Category=Closed-Source, Backbone=GPT-4o2026.0171.971.3271.6271.7673.2273.2173.0769.1471.91
GPT-5.1Model Category=Closed-Source2026.0171.3270.4570.4571.3272.7872.6374.2467.8971.39
RAG + CoTModel Category=Closed-Source, Backbone=GPT-5.1, Reasoning Strategy=RAG + Chain-of-Thought2026.0169.1469.8767.8370.3172.6370.8972.6365.669.86
Self-ConsistencyModel Category=Closed-Source, Backbone=GPT-5.1, Reasoning Strategy=Self-Consistency2026.0168.0868.5166.9670.1271.1871.7271.9762.3968.87
GPT-4oModel Category=Closed-Source2026.0167.3967.3965.9469.1469.7270.671.4761.8667.94
CoTModel Category=Closed-Source, Backbone=GPT-5.1, Reasoning Strategy=Chain-of-Thought2026.0167.161.2863.6170.3172.3470.0171.3260.9967.12
DeepSeek-3.2Model Category=Open-Source2026.0166.8166.8161.286968.767.896951.6765.15
SoTModel Category=Closed-Source, Backbone=GPT-5.1, Reasoning Strategy=Structured-of-Thought2026.0165.2164.6362.4567.6968.8568.2769.7256.1965.37
MED-COREASONERModel Category=Open-Source, Backbone=Qwen3-30B2026.0164.9262.1559.3965.6566.3867.8368.4139.0161.72
Qwen3-30BModel Category=Open-Source2026.0161.8656.9154.1561.7261.7264.4863.3227.2256.42
Qwen2.5-72BModel Category=Open-Source2026.0159.3956.4854.1559.8361.8661.5160.2630.1355.45
Qwen2.5-32BModel Category=Open-Source2026.0158.2254.7348.3359.158.6660.4158.9526.7653.15
Claude-3.5-haikuModel Category=Closed-Source2026.0156.3357.6454.8860.761.7259.9762.5935.8156.21
LLaMA3.1-70BModel Category=Open-Source2026.0151.0948.3347.4559.5360.8460.2660.8444.2554.07