Omni-MATH
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
Omni-MATH
78.5Accuracy
135
Omni-Math
78.5Accuracy
41
Omni-MATH
0.088ECE
28
OMNI-MATH
39.55Overall Accuracy
25
Omni-MATH
28.16Avg@4 Accuracy
18
Omni-MATH
35.4Best-of-N Accuracy
17
Omni-MATH
1,046.36AUTC
17
Omni-MATH
37Algebra Accuracy
16
Omni-Math
25.34Average Score @8
14
Omni-MATH terminal answers
0.929AUROC
11
Omni-MATH
8.67ECE
11
Omni-MATH
0.859AUROC
11
OMNI-MATH Hard (val)
38.1Accuracy
10
OMNI-MATH Medium (val)
61.15Accuracy (Next-token Reasoning)
10
OMNI-MATH Easy (val)
76.89Accuracy
10
Omni-MATH
54.1Score
10
Omni-MATH Dataset C
23.22Score (Reference)
8
Omni-MATH 4,415 problems (Full Set)
64Accuracy
8
Omni-MATH human-annotated Reasoning episodes (gold set)
86.33Accuracy
8
Omni-MATH OOD
75.48HV
6
Omni-MATH Tier 2
42.7Success Rate (SR)
6
Omni-MATH
79.1Correlation
5
Omni-MATH (Dataset U)
15.85Reference Score (S)
4
Omni-MATH Rule 2,821 problems
69.7Accuracy
4
Omni-MATH
32.2Accuracy (Omni-MATH)
4