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

Knowledge-intensive reasoning research benchmarks · SOTA2 Research

Knowledge-intensive reasoning

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

85Avg Score

75

May 20, 2026

87Accuracy

51

Jun 11, 2026

Llama3.1-8B + ARPO

34.8F1 Score

43

Jun 2, 2026

Llama3.1-8B + ARPO

0.654F1 Score

41

Jun 11, 2026

2WikiMultiHopQA

81.5Accuracy

38

Jun 11, 2026

72.7Overall Score

35

Jul 2, 2026

Knowledge-Intensive Reasoning Suite 2Wiki., Bamb., HQA, MuSi., SimQA

Qwen2.5-7B + SFT-then-RL

58.42Wiki Score

25

Apr 13, 2026

Llama3.1-8B + ARPO

73.8F1

23

Jun 11, 2026

33.5WebWalker Accuracy

20

Jun 11, 2026

2wikiMultiHopQA

Qwen2.5-7B + GRPO

76.1F1 Score

18

Feb 27, 2026

Llama3.1-8B + ARPO

30.5F1 Score

18

Feb 27, 2026

87Average Score

18

Feb 26, 2026

60.4F1 Score

15

Jun 2, 2026

2WikiMultihopQA

58.6F1 Score

15

Jun 2, 2026

38.89Result Score

14

Apr 13, 2026

GPQA ambiguity-augmented

42.8Accuracy

11

Apr 28, 2026

0.89Average Score

9

Feb 26, 2026

90.2Score

7

May 13, 2026

MMLU-CF first 1,000 samples (test)

74.2Exact Match Accuracy

7

Feb 26, 2026

Knowledge-intensive reasoning suite (HotpotQA, 2WikiMultihopQA, Musique)

43.6HotpotQA Score

6

Feb 26, 2026

52F1 Score

5

Jun 2, 2026

Generalization Verification

KDCM + Code Module

99.18Hits@1

5

Feb 26, 2026

Popular datasets

2WikiMultiHopQA GPQA MuSiQue Bamboogle WebWalker

Follow for updates

Join SOTA2 Contact