AGIEval
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
AGIEval MATH
95.7Accuracy
99
AGIEval
71.6Accuracy
46
AGIEval (test)
62.3Accuracy
37
AGIEval
70.4Exact Match
33
AGIEval-MATH (test)
93.3Accuracy
31
AGIEval
70.22Accuracy
29
AGIEval English
74.4Score (%)
21
AGIEval
51.05Score
18
AGIEval
58.1Accuracy
16
AGIEval Out-of-Domain Law (test)
43.41Average OOD Accuracy
16
AGIEval en
2.132Speedup Ratio
15
agieval
63.71Accuracy
14
AGIEval
43.92Vanilla Accuracy
14
AGIEval MATH Level-5
64.45Accuracy
13
AGIEval
32.11Accuracy
12
AGIEval
48.88AGIEval Reasoning Accuracy
10
Agieval
37.3Accuracy
8
AGIEval En
78.73CoT EM
8
AGIEval (test)
27AGIEval (3-shot)
8
Agieval Cn
36.58Accuracy
7
AGIEval
30.91Accuracy
6
AGIEval (test)
28.3AQUA-RAT
5
AGIEval G
72Accuracy
4
AGIEval
83.1EM
3
AGIEval 5-shot
27.2LSAT-RC (5-shot)
3