Estimate how much it costs to send your PDF to GPT-4o, Claude, Gemini, or any LLM API. Editable prices, no upload.
Every call to an LLM API charges by token count. When you send a PDF to GPT-4o, Claude, or Gemini via their API, the bill depends on how many tokens your document contains. Knowing the cost before you make the call helps you pick the right model and decide whether to chunk or summarize first.
This tool extracts text from your PDF locally, counts the tokens using the same tokenizer GPT-4 uses (cl100k_base), and multiplies by the published per-token price for each model. You can edit prices or add custom models if your rates differ.
Your PDF content stays on your device. Token counting and cost multiplication happen locally. Nothing is sent to any server or API.
They reflect published list prices as of mid-2026. Actual costs may differ based on volume tiers, commitments, or provider updates. Always verify on the provider's pricing page before committing to a budget.
The estimate shows input cost only (sending the PDF to the model). Output cost depends on how long the model's response is, which varies by prompt and task.
Yes. Click Edit prices and then Add model. Enter the name, provider, price per million tokens, and context window size.