Dialog research benchmarks

Benchmarks

Dataset NameSOTA methodMetricTrendResultsLast Updated

bAbI dialog 1.0 (OOV)

QRN

0.023Avg Error Rate

Feb 26, 2026

WoW (Wizard of Wikipedia) (test)

INFO-RAG

11.38F1 Score

Feb 26, 2026

DSTC2

QRN

0.489Average Error Rate

Feb 26, 2026

bAbI dialog

QRN+

1.5Average Error Rate

Feb 26, 2026

DSTC2 (test)

QRN

48.9Average Error Rate

Feb 26, 2026

bAbI dialog Standard 1.0

QRN

1.5Average Error Rate

Feb 26, 2026

Real-world Dialog Domains Aggregate

DFPO

87.31Average Accuracy

Feb 26, 2026

Financial Services Out-of-Domain

DFPO

84.3Accuracy

Feb 26, 2026

Social & Entertainment Out-of-Domain

Baseline

87.13Accuracy

Feb 26, 2026

Healthcare & Wellness Out-of-Domain

Baseline

90.23Accuracy

Feb 26, 2026

Transportation & Travel Out-of-Domain

DFPO

90.5Accuracy

Feb 26, 2026

Life Services In-Domain

Baseline

86.73Accuracy

Feb 26, 2026

MT-Bench

CoSpec

9.43Score

May 26, 2026