Token Count Estimator
Approximate tokens and cost for any text.
Estimated tokens
54
224 chars · 38 words
Cost as input tokens (one call)
Claude Opus 4.7$0.000810
Claude Sonnet 4.6$0.000162
Claude Haiku 4.5$0.000054
GPT-4.1$0.000270
Gemini 2.5 Pro$0.000135
Estimate uses both the character/4 rule and the word/0.75 rule, averaged. For exact counts, use a real tokenizer (tiktoken for OpenAI, Anthropic's count-tokens API for Claude). This estimate is usually within 5-10% for English text and gets less accurate for code, JSON, or non-Latin languages.
About
Paste any text. Get an estimated token count using both the character/4 and word/0.75 heuristics, plus the cost as input tokens across major LLM providers. For exact counts use the provider's tokenizer.
How to use
- Paste text.
- Read estimated token count.
FAQ
How accurate is this?+
Within 5-10% for English text. Less accurate for code (more tokens per char), JSON (variable), non-Latin scripts (sometimes 2-3x more). For exact counts, use the official tokenizer.