Categories Companies Research

Loading the SOTA2 catalog…

Explore

Categories Companies Research

Contribute

Join SOTA2 Contact

SOTA2

© 2026 SOTA2

Build in Bellevue, USA

WizWand is now SOTA2 Research

Papers Benchmarks Datasets Tasks

Mathematical Reasoning research benchmarks – Page 107 · SOTA2 Research

Mathematical Reasoning

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

GSM8K s (subsampled 2500)

96Accuracy

8

Feb 26, 2026

44.01Avg@32

8

Feb 26, 2026

38.4Accuracy

8

Feb 26, 2026

97.3pass@1

8

Feb 26, 2026

GSM8K zero-shot

Confidence-Adaptive Decoding (CAD)

77.6Accuracy

8

Feb 26, 2026

AIME Think (test)

72.4Accuracy

8

Feb 26, 2026

AIME No-think (test)

62.9Accuracy

8

Feb 26, 2026

43.66Accuracy

8

Feb 26, 2026

79.92Accuracy

8

Feb 26, 2026

0.928Avg@16 Accuracy

8

Feb 26, 2026

Qwen2.5-7B-Instruct (GRPO)

74.7Avg@16 Accuracy

8

Feb 26, 2026

27.3Avg@16 Accuracy

8

Feb 26, 2026

37.1Avg@16 Accuracy

8

Feb 26, 2026

2.71Speedup

8

Feb 26, 2026

MGSM Code Switched P, I - (EN), Q(X) (Avg)

Bloom-7.1B (Full scope SFT)

100Language Consistency

8

Feb 26, 2026

MGSM Monolingual P, I, Q - (X) (Avg)

Bloom-7.1B (Full scope SFT)

100Language Consistency

8

Feb 26, 2026

61Accuracy

8

Feb 26, 2026

GSM8K (evaluation)

93.3Accuracy

8

Feb 26, 2026

79Accuracy

8

Feb 26, 2026

Benchmark LLM (BenchAcc)

89Accuracy

8

Feb 26, 2026

Math Benchmarks (GSM8K, MATH, AMC23, AIME24) (test)

96Accuracy (GSM8K)

8

Feb 26, 2026

NuminaMath (val)

Claude 3.7 Sonnet

21.7Accuracy

8

Feb 26, 2026

43.4Pass@1 Accuracy

8

Feb 26, 2026

MATH-Vision mini (test)

42.11ALG

8

Feb 26, 2026

GSM-Plus (mini)

CRESCENT (Llama3-8B-Instruct)

52.8Accuracy

8

Feb 26, 2026

Page 107 of 152

Popular datasets

AIME MATH GSM8K AMC OlympiadBench

Follow for updates

Join SOTA2 Contact