The Token Tax Index: 28 Languages, Measured
Token counts for the same text in 28 languages under o200k_base. Japanese costs 118 percent more than English using half the characters. Open dataset, CSV and JSON.
Do Non-English Prompts Cost More Tokens? I Measured Eight Languages
The same instruction costs 73 percent more tokens in Japanese than English, using less than half the characters. Measured with the real GPT-5.6 tokeniser, with a tool to try it yourself.
