PromptPriceCalculate

Updated with 2026 AI pricing

Estimate your AI costs before you send the prompt.

Pick a model, paste your question, choose how long the answer might be, and see the likely cost in seconds.

Example receipt

$0.0431

A long answer usually costs more than the question.

$/1M inputOutput costCached tokensBatch pricing

Quick estimate

No API key needed. Pick a model, paste your question, choose the answer length, and get a close cost estimate.

$5/M input$30/M output400,000 ctxCache pricingBatchFlexPriority

3. How long will the AI answer be?

Your estimate will appear here.

Paste a prompt, choose an answer length, and calculate.

Token-to-word reference

1,000 tokens ≈ 750 words of text

This is a close estimate. For exact billing, use expert controls with provider usage JSON or token counts.

Real-world token examples

Short prompt

"What is 2+2?"

Token count~5 tokens

Medium prompt

"Write a 100-word product description for a smart coffee maker."

Token count~25 tokens

Long prompt

"Analyze this article and summarize the main arguments..."

Token count~200+ tokens

Short answer

"Yes, that is correct."

Token count~5 tokens

Medium answer

"Here is a concise explanation with a few helpful details."

Token count~50 tokens

Long answer

"A detailed response with examples, bullets, and context."

Token count~200+ tokens

Why AI answers cost more

The model has to generate answer tokens one by one, so output is often the expensive part.

When exact accuracy matters

Paste API usage JSON or provider token counts in expert mode for the closest billing-grade estimate.

How to lower the cost

Try shorter prompts, shorter answers, cheaper models, batch pricing, and cache-friendly repeated context.

What is an AI token?

A token is a small chunk of text used by AI models for billing and processing.

Why do input and output tokens cost different amounts?

Providers often price generated output higher because generation uses more compute.

Do you store my prompt?

No. The MVP stores token counts, model choice, consent, and cost buckets, not raw prompt or output text.

Is this calculator exact?

It is a close estimate unless you paste API usage JSON or manually enter provider-reported token counts.