Verified October 2026

Home › Tools › Tokens per word: Chinese

Tokens per word in Chinese

Chinese averages about 1.5 tokens per character on current BPE tokenizers, which is 13% more tokens than the same count in English. Counted per character, not per word. Common characters are usually one token; rare ones take two or three.

1.5Tokens per character
0.67characters per token
1,500Tokens per 1,000 characters
+13%vs English

Chinese to tokens table

Characters → Tokens (Chinese, 1.5 tokens per character)
CharactersTokens
100150
500750
1,0001,500
5,0007,500
10,00015,000
50,00075,000
100,000150,000

Language comparison

LanguageTokens per characterTokens for 1,000Cost multiplier vs English
English1.331,3301.00×
Spanish1.71,7001.28×
French1.81,8001.35×
German1.91,9001.43×
Portuguese1.71,7001.28×
Russian2.42,4001.80×
Arabic2.92,9002.18×
Hindi3.23,2002.41×
Chinese1.51,5001.13×
Japanese1.81,8001.35×

Ratios are per word except Chinese and Japanese, which are per character. Comparing them directly is approximate — the honest way to budget is to tokenize a sample of your own corpus.

Cost of 1,000 characters in Chinese

Cost of 1,500 tokens (1,000 characters) — USD, rates verified October 2026
#ModelProviderInputOutputBlended
1GPT-6 LunaOpenAI$0.00015$0.00075$0.00034
2GPT-5.6 LunaOpenAI$0.0003$0.0018$0.00075
3GPT-5.4 nanoOpenAI$0.0003$0.00187$0.00077
4Gemini 3.5 Flash-LiteGoogle Gemini$0.00045$0.00375$0.00139
5DeepSeek V4.1 FlashDeepSeek$0.00045$0.0018$0.0009
6GPT-5.4 miniOpenAI$0.00113$0.00675$0.00281
7Gemini 3.8 FlashGoogle Gemini$0.00113$0.00562$0.00253
8Gemini 3.7 FlashGoogle Gemini$0.00113$0.00562$0.00253
9Claude Haiku 4.5Anthropic$0.0015$0.0075$0.00337
10Gemini 2.5 ProGoogle Gemini$0.00187$0.015$0.00562
11Grok 4.3xAI Grok$0.00187$0.00375$0.00281
12DeepSeek V4 ProDeepSeek$0.00198$0.00594$0.00347
13Gemini 3.5 FlashGoogle Gemini$0.00225$0.0135$0.00562
14GPT-5.2OpenAI$0.00263$0.021$0.00788
15GPT-6.1 SolOpenAI$0.003$0.015$0.00675
16GPT-6 SolOpenAI$0.003$0.015$0.00675
17GPT-5.6 TerraOpenAI$0.003$0.018$0.0075
18GPT-4.1OpenAI$0.003$0.012$0.006
19Claude Sonnet 5Anthropic$0.003$0.015$0.00675
20Gemini 3.1 Pro (Preview)Google Gemini$0.003$0.018$0.0075
21Grok 4.6xAI Grok$0.003$0.009$0.00525
22GPT-5.4OpenAI$0.00375$0.0225$0.00937
23Claude Sonnet 4.6Anthropic$0.0045$0.0225$0.0101
24GPT-5.6 SolOpenAI$0.006$0.03$0.0135
25GPT-5.5OpenAI$0.0075$0.045$0.0187
26Claude Opus 5Anthropic$0.0075$0.0375$0.0169
27GPT-6 AstraOpenAI$0.015$0.075$0.0338
28Claude Fable 5.1Anthropic$0.015$0.075$0.0338

Input = 1,500 prompt tokens. Output = 1,500 generated tokens. Blended = input + 25% of output, the same weighting used across this site.

What this means for a multilingual product

If you serve Chinese and English from the same budget, Chinese consumes 1.1× the tokens for the same number of characters. Either the price per token absorbs that, or the Chinese product quietly runs at a lower margin. Measuring the ratio on your own text is the only reliable way to price it.

FAQ

How many tokens is 1,000 characters in Chinese?

About 1,500 tokens, using 1.5 tokens per character. That is 13% more tokens than the same count in English (1,330).

Why is Chinese more expensive per character?

Counted per character, not per word. Common characters are usually one token; rare ones take two or three. Providers bill per token, not per character, so a language that fragments into more tokens costs more for the same meaning.

How do I get an exact count for Chinese?

Run a sample of your own text through the tokenizer of the model you will be billed on. Public ratios are averages; product names, numbers and formatting in your own data can move the result by 20% or more.

Keep going

Calculate a real monthly bill