Gemini token counter
Google does not publish Gemini's tokenizer, so browser-side tools can only estimate. This counter does both honestly: a live estimate as you type, and one click asks Google's own countTokens API, so the number you see is the number the API bills.
0 words · 0 characters
Gemini models at a glance
Your text counted against every current Gemini model. Counting against every provider at once? Use the full token counter, or see every model's limit in the context-window comparison.
| Model | Your text | Context window | Max output | Input $/M tokens |
|---|---|---|---|---|
| Gemini 3.1 Pro $2 up to 200K-token prompts, $4 beyond | — | 1M ~750K words | 64K | $2 |
| Gemini 2.5 Pro $1.25 up to 200K-token prompts, $2.50 beyond | — | 1.0M ~786K words | 66K | $1.25 |
| Gemini 2.5 Flash | — | 1.0M ~786K words | 66K | $0.3 |
Windows and prices verified against provider documentation on 2026-07-07.
Gemini token questions, answered
How does Gemini count tokens?
With Google's own private tokenizer; roughly four characters per token for English text, per Google's documentation. The only exact number comes from the countTokens API, which is what the exact-count button asks.
Is this count exact?
The live count while you type is an estimate. Click the exact-count button and the counter asks Google's countTokens API directly; that result is exact for the listed Gemini models.
How big is Gemini's context window?
The current Gemini Pro and Flash models take 1,000,000 tokens over the API. The Gemini consumer app gates the window by plan: about 32K on the free tier, 128K on AI Plus, and the full million on AI Pro and Ultra.
Does a million-token window mean I should fill it?
Usually not. Models attend unevenly across very long contexts, so a short, deliberate context tends to beat a full window of everything. The window is capacity; what you choose to put in it is what decides the answer.
Counting tokens is managing context by hand. NodePad is a canvas for doing it on purpose: fork side questions, keep every path, and choose exactly what Gemini (or any other model) sees.
Try NodePad free