How is the per-call cost computed?

Cost = (input tokens ÷ 1,000,000 × input price) + (output tokens ÷ 1,000,000 × output price). Input tokens are estimated from your pasted prompt; output tokens come from your expected-completion field.

Why is output more expensive than input?

Generating tokens is more compute-intensive than reading them, so providers price output 3-5x higher than input on most models. That is why trimming verbose responses saves more than trimming prompts.

Can I trust the preset prices?

They are based on published list prices and clearly editable. Providers change pricing often, so override any row with the current rate from your provider dashboard before budgeting.

Is anything sent to a server?

No. Token estimation and all math run locally in your browser. Your prompt text never leaves the page.

Prompt Cost Calculator

Prompt cost calculator

Paste a real prompt, say how long you expect the answer to be, and instantly see what one API call would cost on GPT-4o, Claude Sonnet, Gemini Flash, and the other major models — side by side. It is the fastest way to answer “which model is cheapest for this exact task?” without spreadsheet math.

How it works

The tool estimates input tokens directly from your pasted prompt using an English-calibrated characters-and-words blend (roughly 1 word ≈ 1.3 tokens). You supply the expected completion tokens. Each model row then computes (input ÷ 1M × input price) + (output ÷ 1M × output price). Because output tokens are billed several times higher than input on most models, the completion length usually dominates the total — which is exactly why comparing models on a real prompt beats comparing headline per-million numbers.

Tips and notes

Every price is editable: if you have a volume discount or the provider changed rates, type the new number and the comparison reranks live. Remember that a “cheap” model that needs longer or retried responses can cost more in practice than a pricier model that answers correctly the first time. For high-volume workloads, multiply the cheapest per-call figure by your monthly request count to see the real budget impact.