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

Language Generation research benchmarks – Page 2 · SOTA2 Research

Language Generation

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

Experimental Setup

1Relative Runtime

8

Feb 26, 2026

WebText (completions)

Coherence Tuning

10.16Perplexity (PPL)

7

Feb 26, 2026

56.18ROUGE-1

6

Mar 12, 2026

KGRec-Music (test)

56.18Rouge-1

6

Mar 12, 2026

300K-example corpus (test)

136.02Generation Perplexity

5

May 15, 2026

TinyStories (test)

9.93Grammar

5

Feb 26, 2026

Synthetic data v1 (test)

8.247NLL

4

Feb 26, 2026

92.94MAUVE

3

Jul 3, 2026

67.34Accuracy

3

Feb 26, 2026

Large-scale model pool 15 LLMs

51.9Accuracy

3

Feb 26, 2026

84.5BLEU-2

3

Feb 26, 2026

Agnostic Statistical Language Learning

Pure DP Generation

1Error Rate

2

Apr 10, 2026

One Billion Word 6-gram

0.74JSD

2

Feb 26, 2026

One Billion Word 4-gram

0.35JSD

2

Feb 26, 2026

Page 2 of 2

Popular datasets

C4 OpenWebText Synthetic Data One Billion Word KGRec

Follow for updates

Join SOTA2 Contact