Paste any text and instantly see the token count and estimated cost for every major LLM side by side.
LLM APIs charge per token — a subword unit roughly equal to 4 characters or 0.75 words in English. The same text produces different token counts across models because each uses a different tokenizer. This tool shows you exact counts (for OpenAI models via tiktoken) and close estimates (for others) alongside real-time cost calculations from current API pricing.
Type or paste text in the input area to see token counts.
Need more detail? Compare all model pricing, estimate monthly API costs, or learn how LLM token pricing works.
A token is the smallest unit of text that an LLM processes. Tokenizers split text into subword pieces using algorithms like Byte-Pair Encoding (BPE). Common words are usually a single token, while rare or long words get split into multiple tokens.
"The cat sat" → ["The", " cat", " sat"] = 3 tokens
"Hamburger" → ["Ham", "bur", "ger"] = 3 tokens
"AI" → ["AI"] = 1 token
Rule of thumb: 1 token ≈ 4 characters ≈ 0.75 words in English. Code and non-Latin scripts may tokenize differently. For the full story, read our guide: How LLM Token Pricing Works.
A token is the basic unit of text that large language models process. Tokens are subword units created by a tokenizer algorithm. In English, one token is roughly 4 characters or 0.75 words. Common words like "the" are single tokens, while longer or rarer words may be split into multiple tokens.
Approximately 1,333 tokens for typical English text. The exact count depends on the model's tokenizer and the vocabulary used — technical jargon and code tend to produce more tokens per word than everyday English.
Yes. Each provider uses a different tokenizer algorithm with a different vocabulary. The same text will produce different token counts in GPT-5, Claude, Gemini, and Llama models. The differences are usually small (within 5-10%) but can affect cost estimates at scale.
Use this free token counter tool to check counts across multiple models instantly. For programmatic use, OpenAI provides tiktoken, Anthropic offers a count_tokens endpoint, and Google provides a CountTokens API. All provider counting endpoints are free to use.
LLM APIs charge per token processed. Both your input (prompt) and the model's output are billed separately, with output tokens typically costing 2-5x more than input tokens. Knowing your token count helps you estimate costs and choose the right model for your budget.
Byte-Pair Encoding (BPE) is the tokenization algorithm used by most modern LLMs including GPT, Claude, and Llama. It breaks text into subword units based on frequency patterns learned during training, balancing vocabulary size with the ability to handle rare and novel words.
Get notified when new models launch, pricing changes, or benchmarks update.
Free. No spam. Unsubscribe anytime. We only store derived location metadata for consent routing.