NuminaMath
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
NuminaMath
60.9Accuracy
39
NuminaMath
78.67Accuracy
20
NuminaMath
62.45Math Accuracy
18
NuminaMath in-distribution LEAN (test)
65.17Average@32
12
NuminaMath
45.96Pass@1
9
NuminaMath (val)
21.7Accuracy
8
NuminaMath
12.8Throughput (RPS)
6
NuminaMath (out-of-domain)
42.22Accuracy
6
NuminaMath-LEAN (total)
51Accuracy
4
NuminaMath LEAN (unsolved)
26Accuracy
4
NuminaMath-LEAN (solved-H)
47Accuracy
4
NuminaMath-LEAN (solved-K)
100Accuracy
4
NuminaMath LEAN (In-domain)
1,707.19Average Token Cost
4
NuminaMath subset of 5,000 samples
73.94Accuracy
3