Multi-turn Chat Evaluation
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
8.58MT-Bench Score
42
Jun 11, 2026
79Binary Accuracy
9
Mar 27, 2026
73Binary Agreement
9
Mar 27, 2026
3.61Acceptance Length
8
May 12, 2026
7.925Score
3
Feb 26, 2026
73.9Winrate
2
Mar 18, 2026
71Winrate
2
Mar 18, 2026