MATH
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
Math ID GSM8k ProofNet
99.2GSM8k Accuracy
28
MATH500
93.5Accuracy
28
MATH
28.9Retention
28
MATH 500
5.67Average Acceptance Length
28
MATH-500
95.52Accuracy
28
MATH 500
95Accuracy
28
Math
0.695AUROC
27
MATH 500
94.8Accuracy
27
MATH 500
87.8Success Rate (pass@4)
27
MATH 500
94.77Accuracy
27
MATH (test)
83.8Pass@1
27
MATH
39.66Score
27
Math 500
91.3Flex Match
27
Math
94.2Math Score
27
MATH
90Pass@1
27
MATH
94.2Accuracy
26
MATH-500 1.0 (test)
86.6Accuracy
26
MATH lighteval
95.6Accuracy
26
MATH 500
95.3Accuracy
26
Math500
87.18Accuracy
26
MATH (test)
2.7Latency (s)
26
MATH 500
92.4Pass@1 Rate
26
MATH 500
98Accuracy
26
MATH ProcessBench (test)
9.6Error Rate
26
Math-500
1,094.6Throughput (TPS)
25