How many tokens is a word?
Token-to-word ratios for GPT-5, GPT-4o, Claude, Gemini and Llama with real-time converter.
Read the guide →Free AI Token Calculator — Count Tokens & Estimate API Costs
Count tokens for GPT-5, GPT-4o, Claude, Gemini, Llama, DeepSeek and 90+ models. All calculations run locally in your browser — your text never leaves your computer.
100% Free
No Upload
90+ Models
Instant Results
AI Token Calculator
Why developers use tokenestimate
Σ Exact tiktoken counts
Uses the same Byte-Pair Encoding tokenizer as OpenAI's servers for accurate results.
$ Cost before you code
Estimate API costs before making requests. Supports custom pricing for any model.
# 100% local & private
All calculations run in your browser. No data uploaded, no tracking, no accounts needed.
⎈ Built for AI workflows
Designed for developers, researchers, and AI engineers who need quick token estimates.
Supported models & pricing
Count tokens and estimate costs for 90+ models across all major AI providers.
OpenAI
Anthropic
Developer guides
AI API pricing in 2026
Complete pricing comparison for all major AI models with cost optimization tips.
Read the guide →How to reduce AI API costs
Practical tips for token optimization, prompt engineering, and cost reduction strategies.
Read the guide →Complete guide to AI tokenization
Understanding BPE, tokenizers, and how different models split text into tokens.
Read the guide →Frequently asked questions
What is an AI token?
A token is the basic unit of text that language models read and generate. For English text, 1 token is roughly 0.7 word (or 4 characters). OpenAI bills all API usage in tokens, so understanding token counts helps you estimate costs.
How many tokens is a word?
On average, one English word equals approximately 1.3 tokens with OpenAI's o200k tokenizer. For Chinese text, one character equals about 0.6-1.2 tokens. The exact ratio depends on the model's tokenizer and the text content.
How accurate is this token calculator?
For OpenAI models, the calculator is exact — it uses the actual tiktoken tokenizer (same as OpenAI's servers). For Claude, Gemini, Llama, and other providers, it uses estimate mode (approximately 4 characters per token), accurate within 5-10%.
How much does GPT-4o cost per 1M tokens?
GPT-4o costs $2.50 per 1M input tokens and $10 per 1M output tokens. GPT-4o mini is much cheaper at $0.15 per 1M input and $0.60 per 1M output. All prices are editable in the calculator.
Which models are supported?
We support 90+ models including all OpenAI models (GPT-5, GPT-4o, o1, o3), Anthropic Claude (4, Opus, Sonnet, Haiku), Google Gemini (Pro, Flash), Meta Llama, DeepSeek, Qwen, Mistral, Kimi, GLM, Grok and more.
Is my text uploaded anywhere?
No. All calculations run 100% locally in your browser using JavaScript. Your prompts, code, and documents never leave your computer. You can even disconnect from the internet and the calculator will still work.
Can I use custom pricing?
Yes! You can edit the price per 1M tokens for any model. This is useful for enterprise pricing, custom fine-tunes, or when prices change. Your custom prices are saved in your browser for future sessions.
Why do token counts differ from word counts?
Tokens are not words. Common words like 'the' may be single tokens, while long words are split into multiple tokens. Whitespace is usually attached to the following token. The tokenizer uses Byte-Pair Encoding (BPE) to split text.
Start counting tokens for free
No signup required. Your text never leaves your browser.
Try the Calculator