Benchmark family

MathArena ArXivMath

A family groups the versions of one benchmark. It is not evaluated and has no applicable models of its own; each of its 8 definitions is resolved separately.

MathematicsRolling evaluationResearch mathematics

Definitions

Compare the newest 4 →

Newest first, following the recorded successor chain. Each definition is resolved against models separately.

BenchmarkMathArena ArXivMath08/2026
Active
Successor of MathArena ArXivMath
Scale
57 problems
Primary metric
Accuracy
Applicable models
13
BenchmarkMathArena ArXivMath06/2026
Active
Successor of MathArena ArXivMath
Scale
48 problems
Primary metric
Accuracy
Applicable models
13
BenchmarkMathArena ArXivMath05/2026
ActiveNewest in family
Successor of MathArena ArXivMath
Scale
40 problems
Primary metric
Accuracy
Applicable models
13
BenchmarkMathArena ArXivMath04/2026
Deprecated
Successor of MathArena ArXivMath
Scale
40 problems
Primary metric
Accuracy
Applicable models
13
BenchmarkMathArena ArXivMath03/2026
Deprecated
Successor of MathArena ArXivMath
Scale
30 problems
Primary metric
Accuracy
Applicable models
13
BenchmarkMathArena ArXivMath02/2026
Deprecated
Successor of MathArena ArXivMath
Scale
32 problems
Primary metric
Accuracy
Applicable models
13
BenchmarkMathArena ArXivMath01/2026
Deprecated
Successor of MathArena ArXivMath
Scale
23 problems
Primary metric
Accuracy
Applicable models
13
BenchmarkMathArena ArXivMath12/2025
Deprecated
Scale
17 problems
Primary metric
Accuracy
Applicable models
13

Family methodology

No family-level methodology recorded.

Across definitions

Tasks: None recorded
Sources: None recorded at family level
Last reviewed: 2026-09-22

Applicability does not mean evaluated; benchmark result/model score data is outside this UI contract.