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

Natural Language Understanding and Reasoning research benchmarks · SOTA2 Research

Natural Language Understanding and Reasoning

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

Standard Suite Zero-shot (PiQA, ARC-E, ARC-C, HellaSwag, WinoGrande, BoolQ)

80.7PiQA Accuracy (Zero-shot)

39

May 19, 2026

LLM Downstream Evaluation Suite (ARC-c, BoolQ, Winogrande, GSM8k, MMLU) zero-shot

65.36ARC-c Accuracy (zero-shot)

20

May 14, 2026

MMLU, ARC-c, HellaSwag, BOOLQ, PIQA, WinoGrande zero-shot

69.72Average Score (Zero-shot)

20

Mar 16, 2026

Standard Downstream Benchmarks Two-Shot (val)

56.86ARC-E Accuracy (Normalized)

11

Feb 26, 2026

General Benchmarks Italian

37.47ARC-C-it

6

Feb 26, 2026

Popular datasets

General Benchmarks Standard Downstream Benchmarks MMLU, ARC-c, HellaSwag, BOOLQ, PIQA, WinoGrande LLM Downstream Evaluation Suite Standard Suite

Follow for updates

Join SOTA2 Contact