Multi-turn conversation
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
85.25Average Score
107
Jun 2, 2026
4.64Speedup
76
May 20, 2026
5.23SR
43
Mar 4, 2026
8.7Conversation Rating (1-10)
41
Mar 10, 2026
77.1Win Rate
36
Jun 2, 2026
7.29MAT Score
14
Feb 26, 2026
39.51R1 (Pairwise)
13
Feb 26, 2026
7.36Accuracy
10
Feb 26, 2026
8.28Accuracy
9
Feb 26, 2026
8.54Accuracy
7
Feb 26, 2026
55.4Mt. Score
6
Feb 26, 2026
74.3MTP Acceptance Rate
5
Jun 11, 2026
3.69SGLang Acceptance Length
4
May 12, 2026
15.9Win Rate (1st Turn)
3
Apr 2, 2026
8.5Grounding Score
3
Feb 26, 2026
82.7Accuracy
2
Feb 26, 2026