List your company

Tokenization on Multilingual code evaluation set

3.64Compression Ratio (CR)SA-BPE
3.55683.57843.63.6216Apr 15, 2026

Evaluation results

MethodLinks
SA-BPETokenizer=SA-BPE, Priority criterion=F · log R · L2026.043.6431,5675.860
BPETokenizer=BPE2026.043.6229,2406.230
PickyBPETokenizer=PickyBPE2026.043.6128,6736.221,620
BoundlessBPETokenizer=BoundlessBPE2026.043.5629,3756.71,680