Language Modeling
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
40.88Bits Per Byte (BPB)
6
May 29, 2026
71.63BPB (%)
6
May 29, 2026
82.14BPB (%)
6
May 29, 2026
88.57BPB Score
6
May 29, 2026
8.199CE Loss
6
Mar 19, 2026
18.289CE Loss
6
Mar 19, 2026
2.36Perplexity
6
Mar 6, 2026
2.23Perplexity
6
Mar 6, 2026
3.19Perplexity
6
Mar 6, 2026
3.21Perplexity
6
Mar 6, 2026
8.24Perplexity
6
Mar 6, 2026
8.84Perplexity
6
Feb 26, 2026
37.91Perplexity (LLaMA-2 7/8B)
6
Feb 26, 2026
7.26Perplexity (LLaMA-2 7B/8B)
6
Feb 26, 2026
54.15Perplexity
6
Feb 26, 2026
2.29Perplexity
6
Mar 25, 2026
2.915PPL
6
Feb 26, 2026
12.4PPL
6
Feb 26, 2026
5.94ELBO
6
Feb 26, 2026
5.61PPL
6
Feb 26, 2026
5.9PPL
6
Feb 26, 2026
0.809Bits Per Byte
6
Feb 26, 2026
19.6Perplexity
6
Feb 26, 2026
52.08Perplexity
6
Feb 26, 2026
3.12PPLX
6
Feb 26, 2026