Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

AGIEval

Benchmarks

Task NameDataset NameSOTA ResultTrend
Mathematical ReasoningAGIEval MATH
Accuracy95.7
99
Natural Language UnderstandingAGIEval
Accuracy71.6
46
Comprehensive ExaminationAGIEval (test)
Accuracy62.3
37
General ReasoningAGIEval
Exact Match70.4
33
Mathematical ReasoningAGIEval-MATH (test)
Accuracy93.3
31
General EvaluationAGIEval
Accuracy70.22
29
ReasoningAGIEval English
Score (%)74.4
21
Human-level Standardized Exam EvaluationAGIEval
Score51.05
18
Foundation Model EvaluationAGIEval
Accuracy58.1
16
Out-of-Domain GeneralizationAGIEval Out-of-Domain Law (test)
Average OOD Accuracy43.41
16
General ReasoningAGIEval en
Speedup Ratio2.132
15
General Reasoningagieval
Accuracy63.71
14
Question AnsweringAGIEval
Vanilla Accuracy43.92
14
Mathematical ProficiencyAGIEval MATH Level-5
Accuracy64.45
13
Question AnsweringAGIEval
Accuracy32.11
12
ReasoningAGIEval
AGIEval Reasoning Accuracy48.88
10
ReasoningAgieval
Accuracy37.3
8
General KnowledgeAGIEval En
CoT EM78.73
8
General Intelligence EvaluationAGIEval (test)
AGIEval (3-shot)27
8
Question AnsweringAgieval Cn
Accuracy36.58
7
Standardized exam solvingAGIEval
Accuracy30.91
6
Question AnsweringAGIEval (test)
AQUA-RAT28.3
5
General Intelligence EvaluationAGIEval G
Accuracy72
4
World KnowledgeAGIEval
EM83.1
3
Standardized Exam ReasoningAGIEval 5-shot
LSAT-RC (5-shot)27.2
3
Showing 25 of 28 rows