Can I count tokens for Claude?

Anthropic does not publish a tokenizer that runs in the browser, so the Sim token counter estimates Claude token counts from character and word length and labels them as estimates. For an exact Claude count, call the Anthropic token counting API with the same messages you plan to send.

Is my text sent anywhere?

No. The Sim token counter tokenizes text in your browser. Pasted text is never sent to Sim or stored in the page URL; only the selected models are kept in the link so you can share a configuration.

How many tokens is 1,000 words?

About 1,300 tokens for typical English prose with OpenAI tokenizers, since one token averages about four characters. The ratio varies with vocabulary, formatting, and language, so paste the actual text for a real count.

How do token counts translate to cost?

Each model bills input and output tokens at its own rate per one million tokens. The Sim token counter multiplies your token count by the input rate of each selected model, switching to the higher long-context rate when the count crosses a model's pricing tier threshold; use the Sim LLM API pricing calculator to add output tokens, prompt caching, and monthly volume across 121 models.

Back to models

Free tool

Token counter

Paste a prompt or document to count its tokens and characters and see what it costs as input to OpenAI, Anthropic Claude, Google Gemini, and other models. Counts are exact for OpenAI models and estimated for the rest, and your text never leaves the browser.

Text

0 characters. Text stays in your browser and is never sent to Sim.

Models

GPT-6 Astra, Claude Fable 5.1 +1

Token count and input cost per model
Model	Tokens	Input cost
GPT-6 Astra OpenAI · Input $10/1M	estimate0	$0
Claude Fable 5.1 Anthropic · Input $10/1M	estimate0	$0
Gemini 3.8 Flash Google · Input $0.75/1M	estimate0	$0

Exact counts use OpenAI's tiktoken encodings. Other providers do not publish a browser tokenizer, so their counts are character-based estimates. Input cost uses list prices and excludes the tokens a provider adds for message formatting.

Text

0 characters. Text stays in your browser and is never sent to Sim.

Models

Model	Tokens	Input cost

How LLM tokenizers count text

A tokenizer splits text into the tokens a model reads, and each provider uses its own vocabulary, so the same text yields different counts on different models.

What a token is

A token is the unit a language model reads and writes: a whole word, part of a word, a punctuation mark, or a run of whitespace. In English, one token averages about four characters, so 1,000 tokens is roughly 750 words. Code, non-English text, and numbers usually take more tokens per character.

Exact counts and estimates

OpenAI publishes its tokenizer encodings, so the Sim token counter loads OpenAI's tiktoken encodings in your browser and counts text exactly as the OpenAI API tokenizes it. Anthropic, Google, and most other providers do not publish a browser tokenizer, so their counts are character-based estimates and are labelled as such.

Why token counts matter

Tokens set both the price of a call and whether it fits the model context window. An agent that resends its system prompt, tool definitions, and conversation history on every turn can multiply a small prompt into a large bill, so counting the tokens in each piece shows where to trim.

Frequently asked questions

Paste your text into the Sim token counter and select an OpenAI model. The count comes from OpenAI's tiktoken encodings (o200k_base for GPT-4o and newer models, cl100k_base for older GPT-4 and GPT-3.5 models), the same tokenizer the OpenAI API uses to bill input.