Token & cost analyzer

Counts run on our server against the real BPE tables. Your text is never sent to a model provider.

Unlock all 44 models

0 characters

Models

3 selected

OpenAI

Anthropic

Google

DeepSeek

xAI

Meta

Mistral

Alibaba

Cohere

~ marks an estimate. Only OpenAI publishes an offline tokenizer, so counts for other vendors are scaled from o200k_base and will differ slightly from their billing.

Expected output tokensRequests per monthProPrompt cache hit rate: 0%Pro

Batch modeProShow tokens

Free tier: 3 models, 10,000 characters, one input at a time. Pro unlocks all 44 models, batches, cost projection and CSV export.

Model prices last reviewed 2026-08-05. Counts for providers other than OpenAI are estimates, because no other vendor publishes an offline tokenizer. Always reconcile against your provider's own billing before committing to a budget.