Quick start
- Enter the request count for the period you care about, such as one day, one month, or one prototype test.
- Enter total input tokens per request, including system instructions, user text, retrieved context, chat history, tool messages, and cached input.
- Enter the cached part of total input. Use 0 when you have no cache-read count.
- Enter average output tokens per request, which is the model response length you expect.
- Enter current standard-input, cached-input, and output prices per 1 million tokens from the same model and service-tier rate card.
Best uses
Best when you have a rough request count, token estimate, current provider price card, and a cache-read count from logs or a realistic test assumption.
- Estimate the monthly cost of an AI support bot, writing helper, or internal tool.
- Measure how much a repeated prompt prefix could change input cost when cache reads are reported.
- Compare two model price cards using the same token and request assumptions.
- Turn a token estimate into a rough budget before building a prototype.
What this calculator is solving
The AI Token Cost Calculator helps you do model-budget math without pretending one price card is permanent. You enter current standard-input, cached-input, and output rates, then see each cost separately plus total and per-request cost.
Match each input label on the calculator to request count, total input tokens, the cached part of that input, output tokens, and current standard-input, cached-input, and output rates for the same model and service tier.
The formula in plain language
In plain language: Uncached input tokens per request = total input tokens - cached input tokens. Uncached input cost = uncached input tokens * requests / 1,000,000 * standard input price. Cached input cost uses the cached-input tokens and cached-input price. Output cost uses output tokens and output price. Total cost = uncached input cost + cached input cost + output cost. The examples on the page are there so you can compare your inputs with a worked example before copying the answer.
For 10,000 requests with 1,200 total input tokens, 800 cached input tokens, and 500 output tokens each, rates of $2 standard input, $0.20 cached input, and $8 output per 1M give $8 uncached input cost, $1.60 cached input cost, and $40 output cost. Total cost is $49.60, or $0.00496 per request.
How to read the answer
Read the total as a planning estimate. The three-way split shows whether fresh prompt content, cached context, or generated answers are driving cost. The cache-rate difference shows how the entered cache-read rate compares with pricing those same tokens at the standard input rate.
- Total cost is the estimated bill for the requests you entered.
- Uncached input, cached input, and output costs are split so you can see which side drives the budget.
- Cost per request is useful when comparing models or deciding whether a feature can scale.
- Cost per 1,000 requests and requests for $100 make a tiny per-request number easier to judge.
Common mistakes to avoid
Most bad AI cost estimates come from stale price cards, mixing per-thousand and per-million units, adding cached input on top of total input, forgetting hidden context, or treating cache writes and batch pricing as automatic savings.
- Do not use old model prices from memory.
- Do not forget that long system prompts, retrieved context, and tool messages can be input tokens too.
- Do not add cached input on top of total input. Cached input is a subset of total input in this calculator.
- Do not use a cache-read price for cache writes or storage unless the provider rate card says they match.
- Do not assume batch rates, free credits, taxes, tool calls, or minimum charges are included.
- Do not compare two models unless the request count and token assumptions are the same.
- Do not treat a rough text token estimate as an exact bill. Check real usage logs once the feature runs.
Quick formula
First subtract cached input from total input to find uncached input. The calculator refuses a cached count larger than total input because that would double-count the request.
Uncached input cost = uncached input tokens * requests / 1,000,000 * standard input rate. Cached input cost and output cost use the same pattern with their matching token counts and rates. Add all three costs for the total.
Example: cached support bot month
Say you expect 10,000 support-bot requests in a month. Each request has 1,200 total input tokens, of which 800 are cache reads, plus 500 output tokens. Use example rates of $2 standard input, $0.20 cached input, and $8 output per 1M.
The uncached side is 4,000,000 tokens and costs $8. Cached input is 8,000,000 tokens and costs $1.60. Output is 5,000,000 tokens and costs $40. Total cost is $49.60, or $0.00496 per request.
- At the standard input rate, those cached tokens would cost $16, so the entered cache-read rate changes the estimate by $14.40.
- If the provider reports no cache hit, set cached input to 0 and price all 1,200 input tokens at the standard rate.
- If traffic doubles and everything else stays the same, the estimated bill doubles.
Use batch rates, not a guessed batch switch
Provider batch programs can use different input, cached-input, and output prices. Some also limit supported models or process jobs asynchronously.
For a batch estimate, copy the current batch rates into the same price fields. This keeps the arithmetic visible and avoids baking a temporary discount into the calculator.
Two more sanity checks
A small prototype with 1,000 requests, 300 input tokens, 150 output tokens, $0.15 input, and $0.60 output per 1M comes out to $0.135 total. That is useful for a tiny test, but it does not predict production traffic.
A long-summary workflow with 2,000 requests, 8,000 input tokens, 700 output tokens, $1.25 input, and $5 output per 1M comes out to $27 total. That shows how long documents can make input cost the main driver.
Where estimates drift
Real bills can move away from the estimate when the model provider changes prices, your app adds hidden system text, users paste longer content, or the feature retries failed requests.
Cache writes, cache storage, long-context tiers, batch jobs, free credits, plan minimums, taxes, image/audio/video tools, retrieval systems, hosting, and monitoring are separate unless you model them with the correct rates or another cost line.
Useful related checks
AI cost planning often starts with token length, then expands into general API pricing and content workflow choices. Use the related tools when you need a rough prompt length, a provider-neutral API cost estimate, or a shorter text sample.
Research and references
These official references help you check current model prices, token counting, and provider-specific cache rules before budgeting.
Worked examples for AI Token Cost Calculator
$8 uncached input + $1.60 cached input + $40 output = $49.60 total, or $0.00496 per request.
$0.045 input + $0.09 output = $0.135 total, about $0.000135 per request.
$20 input + $7 output = $27 total; long prompts drive most of the cost.
FAQ in plain language
When should I use the AI Token Cost Calculator?
Use it when your task matches one of these common needs: Estimate the monthly cost of an AI support bot, writing helper, or internal tool. Measure how much a repeated prompt prefix could change input cost when cache reads are reported. It works best when you already know the measurements, amounts, units, or options the page asks for.
What is the AI Token Cost Calculator doing with my inputs?
In plain language: Uncached input tokens per request = total input tokens - cached input tokens. Uncached input cost = uncached input tokens * requests / 1,000,000 * standard input price. Cached input cost uses the cached-input tokens and cached-input price. Output cost uses output tokens and output price. Total cost = uncached input cost + cached input cost + output cost. The examples on the page are there so you can compare your inputs with a worked example before copying the answer.
What do the main AI Token Cost Calculator inputs mean?
Requests: How many model calls you want to estimate, such as one day, one month, or one product test. Total input tokens per request: All tokens sent to the model, including instructions, prompt text, context, tool messages, and any cached part reported inside that total. Cached input tokens per request: The part of total input served from cache. Enter 0 when you have no cache-read count. Output tokens per request: Tokens generated by the model in each response. Standard input price per 1M tokens: The current uncached input rate for the model and service tier you plan to use. Cached input price per 1M tokens: The current cache-read rate. Cache creation or storage can be a separate charge. Output price per 1M tokens: The current provider rate for one million output tokens. This is often different from the input price.
What should I double-check before trusting the answer?
This is a planning estimate, not a live provider bill. Cached input must be part of the total input count, not extra input added on top. Providers can charge separately for cache writes, cache storage, long context, service tiers, batch jobs, tools, images, audio, credits, and taxes. Use the current rate card and real usage logs for budgets that matter. Also check the unit, scale, mode, and result limit because small input changes can change the answer.
How should I read the AI Token Cost Calculator answer?
Start with total cost, then compare uncached input, cached input, and output cost. Cost per request and cost per 1,000 requests make the scale easier to judge. Requests for $100 is a runway estimate using the same average token counts and rates, not a provider spending limit.
Why does the calculator ask me to enter model prices?
Model prices change and different providers charge different rates for input, output, cached input, fine-tuned models, batch jobs, and special tools. Entering the rate yourself keeps the calculator useful without pretending one price is always current.
How does the calculator count cached input tokens?
Enter cached input as the part of total input that received the cache-read rate. The calculator subtracts that part from total input, prices the uncached remainder at the standard rate, and prices cached input separately. Do not add cached input on top of total input.
Related tools
- Prompt Token EstimatorEstimate prompt tokens from text length with a visible rough range and tokenizer warning.
- API Pricing CalculatorEstimate API usage cost from request count, units per request, unit price, fees, and overhead.
- Text SummarizerSummarize pasted notes into a browser-generated draft.
Keep exploring
If this guide is close but not exact, these links keep you near the same kind of problem.
- AI ToolsBrowse the full category for related tools that help with the same job.
- All free toolsSearch the complete Access Free Tools library by task, category, or tool name.
- All calculator and utility guidesFind more plain-language examples, formulas, mistakes, and result explanations.
- Free calculator resourcesStart here when you are not sure which calculator page fits.
Privacy and copying results
Recent answers stay visible only while you work in the current browser tab. They are not sent to a server.
Use Copy answer when you want to save the inputs and result in notes, homework, a message, or a project list. Check the units, labels, and limits before copying.