Response Generation
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
70.3CEA
30
Apr 10, 2026
35.1BLEU Score
27
Feb 26, 2026
88.1BERTScore
24
Jun 2, 2026
1.43Distinct-3
17
Feb 26, 2026
63.9Accuracy
16
Feb 26, 2026
17.81BLEU-1
16
Apr 10, 2026
1,348Elo
16
Feb 26, 2026
35.4BLEU-2
16
Feb 26, 2026
49.5BLEU-1
13
Apr 23, 2026
33.77BLEU-1
13
Apr 23, 2026
-0.96Reward
13
Feb 26, 2026
3.9Fluency
10
May 11, 2026
3.9NIST-2
9
Feb 26, 2026
20.1BLEU-2
8
Jun 11, 2026
3.89BLEU-2
8
Jun 11, 2026
4.39BLEU-2
8
Jun 11, 2026
35.37PPL
8
Feb 26, 2026
2.669NIST-4
8
Feb 27, 2026
0.924Spearman ρ
7
Jun 2, 2026
72.4BS
7
Apr 28, 2026
1,294Elo Rating
7
Feb 26, 2026
36.82ROUGE-L
6
Apr 14, 2026
39.11ROUGE-L
6
Apr 14, 2026
39.55ROUGE-L
6
Apr 14, 2026
3.67Coherence
6
Apr 10, 2026