AI Cost Calculator

Monthly API cost —

Cost per request: —
Input tokens cost per month: —
Output tokens cost per month: —
Cost over 12 months: —
Tokens per month: —

Where the monthly spend goes

The chart appears with your result

Multiply tokens per request by requests per month, divide by one million and multiply by the price per million tokens, doing input and output separately. A chatbot sending 2,000 input and 500 output tokens, 10,000 times a month at $3 and $15 per million, costs $60 + $75 = $135 a month.

About this tool

Developers, product managers and finance leads budgeting for a large language model API need a number before the first invoice arrives. Providers bill by the token, with input (your prompt and context) and output (the reply) priced separately per million tokens, and output usually costs several times more. Enter the typical size of one request, how many you expect per month and the two prices from your provider's pricing page; the tool returns the monthly and yearly spend, the cost of a single call and how the bill divides between input and output, drawn as a donut. It does not know any provider's current price list, so the figures are only as current as the prices you type, and it ignores discounts such as cached prompts or batch processing.

How to use it

Size one request
Enter the average input tokens (prompt plus context) and output tokens (the reply) for a typical call.
Set the volume
Enter how many requests you expect in a month across all users.
Enter the prices
Type the input and output prices per million tokens from the provider's pricing page for the model you plan to use.
Read the split
Compare the input and output costs and the donut to see which side of the request is worth trimming.

Examples

1M input + 1M output at $3 / $15

Input tokens per request: 1000000
Output tokens per request: 1000000
Requests per month: 1
Input price per million tokens: 3
Output price per million tokens: 15

Result Monthly API cost: $18.00
Cost per request: 18
Input tokens cost per month: $3.00
Output tokens cost per month: $15.00
Cost over 12 months: $216.00
Tokens per month: 2,000,000

Input cost = 1000000 × 1 ÷ 1,000,000 × $3 = $3.00
Output cost = 1000000 × 1 ÷ 1,000,000 × $15 = $15.00
Monthly cost = $3.00 + $15.00 = $18.00
Per request = $18.00 ÷ 1 = $18

Per-million pricing read literally: one million tokens in each direction costs the two list prices added together.

10,000 chats of 2,000 in / 500 out

Input tokens per request: 2000
Output tokens per request: 500
Requests per month: 10000
Input price per million tokens: 3
Output price per million tokens: 15

Result Monthly API cost: $135.00
Cost per request: 0.0135
Input tokens cost per month: $60.00
Output tokens cost per month: $75.00
Cost over 12 months: $1,620.00
Tokens per month: 25,000,000

Input cost = 2000 × 10000 ÷ 1,000,000 × $3 = $60.00
Output cost = 500 × 10000 ÷ 1,000,000 × $15 = $75.00
Monthly cost = $60.00 + $75.00 = $135.00
Per request = $135.00 ÷ 10000 = $0.0135

20 million input tokens cost $60 and 5 million output tokens cost $75. Output is a fifth of the volume but more than half of the bill.

How it is calculated

Cost = (Tin × N ÷ 1,000,000 × Pin) + (Tout × N ÷ 1,000,000 × Pout)

Tin: input tokens per request
Tout: output tokens per request
N: requests per month
Pin: price per million input tokens
Pout: price per million output tokens

Token-billed APIs charge each direction on its own meter. Total monthly tokens on each side are the per-request count times the request volume; dividing by one million converts them to the unit prices are quoted in. The two costs are added for the month, multiplied by 12 for the year, and divided by N for the cost of one call.

When not to use it

For flat-fee subscriptions such as a chat app seat, which are not billed per token.
For image, audio or video models priced per image, second or minute.
When most of your prompt is served from a cache billed at a reduced price.

Common mistakes

Entering the price per thousand tokens where the field asks for the price per million.
Counting only the user's message and forgetting the system prompt and conversation history resent on every turn.
Swapping the input and output prices, which can understate the bill several times over.
Estimating tokens from characters in languages other than English, where one word often takes more tokens.

Frequently asked questions

How many tokens is a word?

For English text, a common rule of thumb is about 0.75 words per token, so 1,000 tokens is roughly 750 words. Code, numbers and many non-English scripts split into more tokens per word, so measure a sample with your provider's tokenizer when the budget is tight.

Why is output priced higher than input?

Generating each output token takes a full pass through the model, one token after another, while the input can be processed in parallel. Providers pass that difference on, typically charging three to five times more per output token.

Does a chat conversation get more expensive with each turn?

Yes, if the full history is sent with every message. Turn ten resends the nine earlier exchanges as input, so input tokens per request grow with the conversation. Use the average input size across all turns, not the size of the first message.

What do cached or batch discounts change?

Some providers bill repeated prompt prefixes at a reduced cached price and offer lower prices for batch jobs that can wait hours. This tool uses one price per direction, so enter a blended price if a known share of your traffic gets the discount.

Where do I find the prices to enter?

Each provider publishes a pricing page listing input and output prices per million tokens for every model. Prices change often and differ between model sizes, so copy them on the day you build the estimate.

How do I estimate requests per month?

Multiply expected daily active users by the average calls each makes per day, then by about 30. For a background job, count documents processed per run times runs per month.

Cite this tool

Using this tool in an assignment, article or report? Copy a citation, or a link for your website.

APA

Tools Directory Editorial Team. (2026, October 25). AI Cost Calculator. Tools Directory. Retrieved October 27, 2026, from https://tools.milunir.com/text-tools/ai-writing-assistants/ai-cost-calculator/

MLA

Tools Directory Editorial Team. “AI Cost Calculator.” Tools Directory, 25 Oct. 2026, https://tools.milunir.com/text-tools/ai-writing-assistants/ai-cost-calculator/. Accessed 27 Oct. 2026.

Chicago

Tools Directory Editorial Team. “AI Cost Calculator.” Tools Directory. Last modified October 25, 2026. Accessed October 27, 2026. https://tools.milunir.com/text-tools/ai-writing-assistants/ai-cost-calculator/.

Link

<a href="https://tools.milunir.com/text-tools/ai-writing-assistants/ai-cost-calculator/">AI Cost Calculator – Tools Directory</a>