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

Open-ended generation research benchmarks · SOTA2 Research

Open-ended generation

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

Creative Writing Evaluation Prompts

8.12Average Judge Score

108

Feb 26, 2026

648Win Rate

49

Jul 7, 2026

70.13BLEURT Score

48

Apr 20, 2026

24.3ROUGE-L

40

Feb 26, 2026

TruthfulQA Without Rejected Samples open-ended (full)

74.67Truthfulness

39

Feb 26, 2026

TruthfulQA With All Samples open-ended (full)

82.75Truthfulness

39

Feb 26, 2026

5.18ECE

37

Jun 2, 2026

MLLMU-Bench (Retain Set)

53.1ROUGE-L

30

Jun 2, 2026

MLLMU-Bench (test)

34.5ROUGE-L

30

Jun 2, 2026

0.479WildBench

26

Mar 9, 2026

Wikitext-103 (test)

0.96MAUVE

26

Feb 26, 2026

7,904Win Rate

23

Feb 26, 2026

84.3GPT-4 Score

21

May 29, 2026

19.75Score (Seen)

21

May 26, 2026

5.21ECE

21

Apr 8, 2026

99.1Skywork Reward V2 Score

18

Apr 23, 2026

0.961Skywork Reward V2 Score

18

Apr 23, 2026

HelloBench (HB)

84HB-A Score

17

Apr 14, 2026

WildBench (test)

64.4WildBench Score

17

Mar 9, 2026

TruthfulQA Open-ended

99.6True Score

16

Feb 26, 2026

45.55MM-Vet Score

14

May 20, 2026

LLaVA-Bench In-the-Wild

109.3Score

14

May 20, 2026

84.6Score

14

Feb 26, 2026

NQ entity-swapped (test)

73.73Exact Match

12

Feb 26, 2026

XSum 1,000 samples (test)

23.11ROUGE-L

12

Feb 26, 2026

Page 1 of 3

PreviousNext

Popular datasets

TruthfulQA LLaVA-Bench NQ AlpacaEval MATH

Follow for updates

Join SOTA2 Contact