Free tools/LLM Token Counter & Cost

LLM token counter

Counted in your browser · text input only

Count tokens in your text and estimate input costs across the models in the table.

Tokens
18
Words
16
Characters
93

Cost of this text as input, per model

ModelAs inputFits context?
GPT-6 Luna OpenAI$0.00000✓
Claude Haiku 5.5 Anthropic$0.00000✓
Gemini 3.8 Flash Google$0.00001✓
GPT-6.1 Sol OpenAI$0.00004✓
Claude Sonnet 5.5 Anthropic$0.00004✓
Claude Opus 5.5 Anthropic$0.00007✓

Counts use the installed gpt-tokenizer encoding. Treat them as estimates across these models; provider tokenizers differ. Costs use the 2026-11-12 Standard text-rate snapshot and documented long-prompt tiers. Reserve capacity for system instructions and output.

How to use it

  1. Paste a sample

    Use a representative prompt or document. Include the instructions and context you expect to send with it.

  2. Read the count

    The tokenizer runs in your browser. Words, characters, and tokens are different measures.

  3. Use costs as estimates

    Other models may use different tokenizers. The table also excludes reply tokens and any additional request content.

More questions

Is this the exact token count for every model?

No. The counter uses one tokenizer; other providers can count the same text differently. Use the relevant provider’s tokenizer for an exact request budget.

Does fitting the context limit guarantee a good answer?

No. Leave room for instructions, tools, and the reply, and test whether the model uses the relevant parts of the text correctly.

Method and example

The browser counts text with the installed gpt-tokenizer encoding, then estimates input cost using the reviewed price table.

Illustrative example: If the counter estimates 1,000 tokens and the selected input price is $2 per million, the estimated input cost is $0.002.

Tokenisation differs by model and language. Treat the count as an estimate for this comparison, and use the provider’s tokenizer or returned usage for billing.

More free AI tools.

Discuss your result with us

Bring the result and one workflow to a free 30-minute call. We will review the assumptions and discuss a practical next step.