Natural language generation
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
89.94ROUGE-L
100
May 12, 2026
70.6BLEU
58
Feb 26, 2026
28.6BLEU
18
Feb 26, 2026
60.34GSM8K
17
Jun 12, 2026
0.472METEOR
17
Apr 10, 2026
49.8BLEU
17
Feb 26, 2026
0.39BLEU-1
14
Jun 1, 2026
40.2BLEU-1
14
Jun 1, 2026
0.462BLEU-1
14
Jun 1, 2026
55.27BLEU
13
Feb 26, 2026
0.539BLEU-1
12
Feb 26, 2026
0.733BLEU-1
12
Feb 26, 2026
3.42EER
11
Feb 26, 2026
31.4spBLEU
11
Feb 26, 2026
55.3BLEU
11
Feb 26, 2026
57.7BLEU
11
Feb 26, 2026
65.3BLEU
11
Feb 26, 2026
90Inform
10
Feb 26, 2026
80.8Exact Match Web Precision
9
Feb 26, 2026
53.5ROUGE-2
9
Feb 27, 2026
45.8ROUGE-2
9
Feb 27, 2026
74.23ROUGE-1
8
Mar 17, 2026
52.85ROUGE-1
8
Mar 17, 2026
55.2ROUGE-1
8
Mar 17, 2026
88.84ROUGE-1
8
Mar 17, 2026