MedQA
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
MedQA TW
88.89Accuracy
12
MedQA
79.3Accuracy
12
MedQA
21.05P@1
12
MedQA
61.1EM
12
MedQA
88.76Accuracy
12
MedQA
92.1Accuracy
12
MedQA
84.83MedQA Accuracy
12
MedQA-de (test)
70.2Accuracy
11
MedQA
1,742Token Cost
11
MedQA
0.944Diversity
11
MedQA (first 300 questions)
44.55Average Accuracy
10
MedQA
66.61Accuracy
10
MedQA
71Pass@1
10
MedQA
0.843Accuracy
10
MedQA
4.32Tau (τ)
10
MedQA USMLE
4.34Kendall's Tau (τ)
10
MedQA Medical domain (test)
69.3Success Rate
10
MedQA Chinese United States Medical Licensing Examination (test)
42.77Accuracy
10
MedQA-USMLE Medical Non-MCQ
22.6Accuracy
9
MedQA-USMLE Medical (val)
83Accuracy+
9
MedQA
80.7Accuracy
9
MedQA
13.5Exact Match (EM)
9
MedQA
100AR (%)
9
MedQA In-Distribution
92.19Accuracy
9
MedQA
0ASR
9