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

Dialogue Generation research benchmarks – Page 4 · SOTA2 Research

Dialogue Generation

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

Navigation (test)

Dialogue State Representation framework

4.21Correctness

3

Feb 26, 2026

100AU

2

Feb 26, 2026

CDConv out-of-domain (test)

0.749Distinct-2

2

Feb 26, 2026

BaikeQA out-of-domain (test)

10.04BLEU-4

2

Feb 26, 2026

Human Interactive Experiments

0.86Embodiment Score

2

Feb 26, 2026

Schema-Guided Dialog (test)

44.1ROUGE-2

2

Feb 26, 2026

Human Evaluation 100 dialogue contexts (test)

DialogVED-Sampling

19Fluency (Win)

2

Feb 26, 2026

Recursive Narrative Bank

90Proportion

1

May 20, 2026

49.51Relevance Win %

1

Feb 26, 2026

Wizard of Wikipedia (WoW) (test seen)

41.34Relevance Win Rate

1

Feb 26, 2026

Dialogue (test)

—

—Primary metric

0

Feb 26, 2026

Dailydialog Human Evaluation

—

—Attribute Relevancy

0

Feb 26, 2026

—

—Primary metric

0

Feb 26, 2026

Page 4 of 4

Popular datasets

DailyDialog DuRecDial Wizard of Wikipedia Persona-Chat ConvAI2

Follow for updates

Join SOTA2 Contact