Language Modeling
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
51.98gPPL
14
Feb 26, 2026
2.151Train Loss (PT)
14
May 22, 2026
3.9Language Score (ar)
14
Feb 26, 2026
15.022-7 Score
14
Feb 26, 2026
39.74Perplexity
14
Apr 15, 2026
52.09Perplexity
14
Apr 15, 2026
1.841Loss
14
Feb 26, 2026
88Diversity (div)
14
Feb 26, 2026
0.87Diversity (Div)
14
Feb 26, 2026
1.22Bits Per Token
14
Feb 26, 2026
3.49Bits per Token
14
Feb 26, 2026
0.935Bits Per Token
14
Feb 26, 2026
56.5Perplexity (PPL)
14
Feb 26, 2026
28Test Perplexity (10 epochs)
14
Feb 26, 2026
16.7Validation Perplexity (PPL)
13
Jun 2, 2026
2Bits Per Character (BPC)
13
May 28, 2026
30.39Perplexity
13
May 7, 2026
1.796Loss (Context 8192)
13
May 7, 2026
10.2Perplexity (WikiEn)
13
Mar 24, 2026
54.9Avg-SK Score
13
Mar 24, 2026
64.5Avg-CK Score
13
Mar 24, 2026
26.756Perplexity
13
Jun 1, 2026
2.62CE Loss
13
Mar 13, 2026
43.91Perplexity
13
Feb 26, 2026
26.06Perplexity
13
Feb 26, 2026