Text Generation
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
4.91Importance
20
Feb 26, 2026
2.01Speedup vs AR
19
May 27, 2026
0.456GPT-2 GM Score
19
Mar 23, 2026
32.9Average Score
18
Feb 26, 2026
63.69BLEU
18
Feb 26, 2026
0.658BLEU-4
17
Mar 30, 2026
41ROUGE-L
17
Jun 17, 2026
7.12Helpful Assistant Score
16
Jun 30, 2026
31.19ROUGE-1
16
Mar 9, 2026
320Decoding Throughput (tokens/s)
16
Feb 26, 2026
25.37Delta Precision (δ-P)
16
Feb 26, 2026
100ASR
16
Feb 26, 2026
82.7ASR
16
Feb 26, 2026
77ASR
16
Feb 26, 2026
36.1ROUGE-1
15
May 14, 2026
49.5ROUGE-1
15
May 14, 2026
34.6ROUGE-1
15
May 14, 2026
0.391ROUGE-1
15
May 14, 2026
39.1ROUGE-1
15
May 14, 2026
3.69Amortized per-token Latency (ms)
15
Apr 24, 2026
84.97Average Score
15
Apr 21, 2026
0.724BERTScore
15
Apr 16, 2026
0.006Cost per 1K requests ($)
15
Feb 26, 2026
5.68Speedup Ratio
15
Jun 2, 2026
8.02BLEU
15
Feb 26, 2026