πͺ AI Token Counter & Cost Calculator
π Seguro e Cliente-side
Estimate the token count of any text and the per-model API cost across GPT-4o, Claude Opus 4.7, Gemini 2.5, and more. Heuristic count (β4 chars/token English, β1.5 chars/token CJK) β exact tokenization needs the provider API.
Characters
0
Words
0
Est. tokens
0
Lines
0
(used for output cost column)
Cost per model
| Model | Input $/1M | Output $/1M | Input cost | Output cost | Total |
|---|
How to use
- Paste the prompt, document, or transcript you want to estimate.
- The character/word/token/line counts update live.
- Adjust the estimated output tokens if you expect a long response (default 500).
- The cost table shows input cost, output cost, and total per popular model.
Token counting accuracy: this tool uses a fast character-based heuristic that detects CJK characters (Korean, Japanese, Chinese) and adjusts the chars-per-token ratio. Real BPE tokenization (cl100k_base, o200k_base, etc.) requires the provider API or a 1-3MB vocabulary download. For cost estimation purposes the heuristic is typically within Β±10% of the exact count.
Pricing: based on publicly listed per-million-token prices as of 2026-05. Always check the provider's pricing page before billing-critical decisions.