Share your thoughts, 1 month free Claude Pro on usSee more

Tokenization Efficiency on Gujarati (gu)

4.69Bytes per Token

LLaMA-4

Updated 25d ago

Evaluation Results

Method	Links
LLaMA-4 2025.11		4.69
Gemma-3 2025.11		6.31
GPT-oss 2025.11		6.68
Sutra 2025.11		7.36
Sarvam-2B 2025.11		7.62
MUTANT-Indic 2025.11		8.95