AI Prompt Character & Token Limits: ChatGPT, Claude & Gemini Compared

AI models don't measure your prompt in characters or words — they measure it in tokens, a unit that doesn't map cleanly to either. Here's what tokens actually are, how context windows compare across ChatGPT, Claude and Gemini, and how to estimate your prompt length.

What Is a Token?

A token is the basic unit a language model processes text in — roughly a word, part of a longer word, or a punctuation mark, rather than a single character. As a rough approximation for English text, one token is about 4 characters or about 0.75 words. So a 1,000-word prompt is roughly 1,300 tokens, and a 1,000-character prompt is roughly 250 tokens.

Context Window Comparison by Model (2026)

For the complete interactive reference with all fields for every platform, visit our all-platforms character limit cheat sheet — with 16 platforms in one page and links to dedicated tools for each.

Why Token Count Isn't the Same as Character or Word Count

Tokenization doesn't split cleanly on whitespace or fixed character counts — common words are often a single token, while rare words, made-up terms, or non-English text can split into several tokens each. This is why the "4 characters per token" rule is only an approximation: a prompt full of common English words will tokenize more efficiently (fewer tokens per character) than one full of technical jargon, code, or non-English text.

Context Window Includes the Whole Conversation, Not Just Your Prompt

A model's context window covers the entire conversation — your prompt, any attached documents, the model's previous responses, and its upcoming response — not just the single message you're typing. A long conversation history or a large pasted document can consume most of the available context window before you've asked your actual question.

Practical Guidance for Long Prompts

• If you're pasting a long document, check its rough character count first — dividing by 4 gives a quick token estimate
• Code and non-English text typically tokenize less efficiently than plain English prose, using more tokens per character
• If you're near a model's context limit, summarizing earlier conversation turns or removing unnecessary pasted content frees up room for the model's response
• Different models tokenize differently — the same prompt can use a meaningfully different token count depending on which model processes it

See our character counter vs word counter guide for the broader distinction between these measurement units, and check your prompt length directly with our AI prompt length checker.