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

Task-oriented Dialogue research benchmarks – Page 2 · SOTA2 Research

Task-oriented Dialogue

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

88.3Success Rate

9

Apr 21, 2026

SMCalFlow (test)

Noisy channel online decoding

64.29SacreBLEU

9

Feb 26, 2026

DP-SGD (no replay)

20.6Backward Transfer (BWT)

8

Jun 2, 2026

68F1 Score

7

Apr 21, 2026

τ-bench 157 scenarios

45.5Collaboration SR

7

Mar 4, 2026

81.29Inform Rate

6

Feb 26, 2026

Real-world Dialogue Benchmark (combined)

72.5Life Services TC

6

Feb 26, 2026

MultiWOZ 89 scenarios

92.7Collaboration SR

5

Mar 4, 2026

MultiWOZ (Human Evaluation)

91.67Success Rate

5

Feb 26, 2026

67.88Micro F1

4

Feb 26, 2026

86.3Inform Rate

4

Feb 26, 2026

MultiWOZ Restaurant domain 2.1

84Inform Rate

4

Feb 26, 2026

MultiWOZ Hotel domain 2.1

75Inform Rate

4

Feb 26, 2026

MultiWOZ 2.1 (train)

0.813Inform Rate

4

Feb 26, 2026

MultiWOZ Attraction domain 2.1

86Inform Rate

4

Feb 26, 2026

In-car personal assistant dataset realtime dialogues

3.98Fluency

4

Feb 26, 2026

MultiWOZ ConvLab simulator

89.4Completion Rate

3

Feb 26, 2026

DSTC9 shared task Human Evaluation (test)

74.8Avg Success Rate

3

Feb 26, 2026

In-car personal assistant dialogue dataset (test)

3.96Correctness

3

Feb 26, 2026

Human Evaluation 400 dialogues

74DS Success Rate

2

Feb 26, 2026

Page 2 of 2

Popular datasets

MultiWOZ SGD StarV2 Human Evaluation KVRET

Follow for updates

Join SOTA2 Contact