We use cookies to keep you signed in. Cookie policy
We use cookies
CalculatorAI uses cookies to keep you signed in and remember your preferences. Optional cookies help us understand how the site is used. Cookie policy
What ChatGPT, Claude, Gemini and Other Models Really Cost
Total cost
One request to GPT-5.6 Terra
Input tokens
150
$0.000300
Output tokens
500
$0.006000
If you keep this up
What drives your monthly bill
What you send
$0.18255%
What it writes back
$3.6595%
The AI's answers are 95% of your bill. Asking for shorter answers is usually the fastest way to spend less.
100 requests
$0.63
1,000 requests
$6.30
10,000 requests
$63.00
Drop in a paystub, bank statement or screenshot — the AI reads it and fills these fields for you.
Income & Expense Tracker is where this stops being a one-off number: keep your real balances there and the projection updates itself.
Describe an invoice and the AI writes it. PDF without watermarks, free.
Most people meet AI pricing as a table of dollars per million tokens and stop there. The table is the easy part. What actually moves a monthly bill is how much text you send, how much you ask the model to write back, and how much of what you send is repeated often enough to be served from cache. Every rate in this calculator comes from the provider's own page — OpenAI, Anthropic, Google, xAI, DeepSeek, Mistral — and is dated on screen.
Output costs two to six times what input costs at every provider here. A prompt asking for a one-paragraph answer instead of a page is usually a bigger saving than switching model.
A chat resends the whole conversation every turn, and an agent resends its instructions on every step. That repetition is where cached-input pricing pays for itself — roughly a tenth of the normal input rate.
Google and xAI reprice a request once the prompt crosses about 200,000 tokens, and the higher rate applies to the whole request. Staying just under the line can halve a call.
The cheapest model in this list is fifty times cheaper than the most expensive. For classification, extraction and routine drafting the gap in output quality is often far smaller than the gap in price.
Estimate your usage honestly, price it per request, then multiply. A monthly figure you can compare against a subscription is worth more than an exact token count you had to guess at.