List your company

Tokenization on Taiwan lexicon

1.7Weighted CostTAIDE
1.59562.30033.0053.7097Jul 7, 2026

Evaluation results

MethodLinks
TAIDEVocab.=318,0802026.071.720
PangolinTokenizerVocab.=114,8222026.072.4827
Qwen 3.6Vocab.=248,0702026.072.5927
Gemma 3/4Vocab.=262,1442026.072.9233
Qwen 2.5/3Vocab.=151,6692026.073.0847
GPT-4oVocab.=200,0192026.073.4540
LLaMA 3.1Vocab.=128,2562026.073.4753
VoxCPM2Vocab.=122,7532026.074.3160