Tiktokenizer

Google Gemini · Local text estimate

Gemini 3.5 Flash Token Counter

Gemini 3.5 Flash uses Google's tokenizer. Use the Gemini API countTokens method for an exact model-specific count; the local tool below uses the documented approximate character ratio for planning, without an API key.

Accuracy note

Estimated text tokens only: characters ÷ 4, rounded up, based on Google's approximate text ratio. This is not a model-exact count or billing guarantee; multilingual text, code and emoji can differ substantially. Use Gemini API countTokens for model-aware counting.

Official counting docs →

Frequently asked questions

How do I count Gemini 3.5 Flash tokens exactly?

Call Google's countTokens method with the same model and content structure you plan to send to generateContent.

Is the local count exact for Gemini?

No. The local result is a character-based estimate and must not be treated as Gemini billing usage.

What is the Gemini 3.5 Flash context window?

Google documents an input limit of 1,048,576 tokens and an output limit of 65,536 tokens.