Language Modeling
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
13.8Perplexity
7
Feb 26, 2026
95.9Perplexity
7
Feb 26, 2026
13.13Perplexity (PPL)
7
Feb 26, 2026
66.9PPL
7
Feb 26, 2026
6.01Subword Perplexity
7
Feb 26, 2026
78.3Perplexity
7
Feb 26, 2026
2.74LM Loss
7
Feb 26, 2026
56.98Accuracy (dep)
7
Feb 26, 2026
11.2PPL
6
Jun 29, 2026
16.69Validation Perplexity (PPL)
6
Jun 29, 2026
1.32Best CE
6
Jun 24, 2026
0.436Bits Per Token (bpt)
6
Jun 24, 2026
30.5Perplexity
6
Jun 9, 2026
5.5Perplexity
6
Jun 9, 2026
0KL Divergence
6
Jun 9, 2026
8.89Perplexity (PPL)
6
Jun 5, 2026
80.2C-Eval Score
6
Jun 5, 2026
1.128BPC
6
Jun 5, 2026
11.46Arxiv Perplexity
6
Jun 4, 2026
16.85Perplexity (Arxiv)
6
Jun 4, 2026
28.56Best Perplexity (PPL)
6
Jun 4, 2026
3.94Forward BPT
6
Jun 2, 2026
1.412Validation Loss
6
Jun 2, 2026
66.71BPB
6
May 29, 2026
65.7BPB
6
May 29, 2026