LLM Token Counter
Processed locallyAI EraCount tokens for GPT-5, GPT-4.1, Claude 4, Gemini 2.5 and more. Supports multiple models.
About LLM Token Counter
The LLM Token Counter estimates token counts for GPT-5, GPT-4.1, Claude 4, Gemini 2.5, and other models. Uses the same tokenisation as OpenAI (tiktoken) via js-tiktoken, so counts match the API for GPT models. Claude and Gemini counts are approximate but useful for budgeting. Tokenisation runs in a Web Worker using WASM—your text never leaves your device. No API calls. Compare multiple models in a table to see how the same text tokenises differently. Stay within context windows and plan your prompts. Essential for prompt engineers, API integrators, and anyone optimising LLM usage.