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

LLM Evaluation research benchmarks · SOTA2 Research

LLM Evaluation

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

Qwen3-1.7B Evaluation Suite (avg)

58.64Average Performance

38

May 7, 2026

51.06AlpacaE

16

Mar 11, 2026

51.32LC Win Rate

16

Jul 7, 2026

JudgeBench (test)

Skywork-Reward-V2-Llama-3.1-8B-40M

79.9Knowledge

16

Mar 4, 2026

Curated Population (MATH-500, MMLU-Redux, SimpleQA)

82.57Accuracy

15

Mar 6, 2026

HuggingFace Open LLM Leaderboard Old (test)

92.08GSM8K Score

14

May 11, 2026

Qwen3-8B + CE-RM-4B

18.2Score

14

Feb 26, 2026

LLM Evaluation Suite (MMLU, IFEval, GSM8K, MATH, HumanEval, MBPP, Hellaswag, GPQA)

51.94MMLU

13

Jul 2, 2026

MILE-RefHumEval

78.98Accuracy

12

Feb 26, 2026

HealthBench (test)

62.6HealthBench Score (%)

11

Mar 26, 2026

Shared (evaluation)

78Tie-aware Accuracy

10

May 29, 2026

Arena-Hard v0.1

Qwen3-8B + CE-RM-4B

78.3Arena-Hard Score

9

Feb 26, 2026

Chinese FuseEval

56.77Win Rate

7

Mar 9, 2026

FuseEval English

55.46Win Rate

7

Mar 9, 2026

64.9Quality Score

6

Jun 4, 2026

Arena-Hard v2.0

65.8Quality

6

Jun 4, 2026

84.3Quality Score

6

Jun 4, 2026

Popular datasets

AlpacaEval Arena-Hard LLM Evaluation Suite JudgeBench WildBench

Follow for updates

Join SOTA2 Contact