Monthly LLM Spend Projector

AI CALCULATOR LLM
Share:
By Checked against sources by , Founder & Architect, RECATOOLS Published Updated Last reviewed

Project your monthly and annual LLM API spend from requests per day and token sizes. GPT-5, Claude, Gemini and more. Free, runs in your browser.

RT-AI-006 · AI Tools
Your browser is too old for this tool’s interactive features. Please update to Safari 13.1, Chrome or Edge 80, Firefox 74 or newer.

Monthly LLM Spend Projector

$0.00 Projected monthly spend
$0.00 Projected annual spend
$0.00 Per day
0 Requests / month
📅 Research current as of 02 Nov 2026 · Sources: List rates from each vendor's own pricing page, through the site's shared rate card. The dates of the oldest and newest row check are printed under the model selector.
Standards, published formulas, and reference data are revised over time — check against a current authoritative source before relying on these figures.
Advertisement
After tool · AD-W1Responsive · Post-tool

How to Use the Spend Projector

Pick your model

Choose the model you'll run in production. Its input and output rates appear under the dropdown.

Estimate daily volume

Enter how many requests you expect per day, plus the typical input and output tokens for one request.

Read the forecast

The tool projects your daily, monthly and annual spend using a year of 365.25 days, so a month is about 30.44 days, plus your monthly request count.

Stress-test the plan

Enter a monthly growth rate to see twelve months laid out, or switch model and watch the figures move. If a cheaper model on the rate card could take the same request, the tool shows what it would cost.

Advertisement
After how-to · AD-W2Responsive

Turning a Per-Request Cost Into a Real Budget

Why daily volume is the number that bites

A single LLM call costs little enough that teams underestimate the bill. Depending on the model and the prompt it can be well under a cent or a few cents, and the tool shows the figure for yours. The cost only becomes real once you multiply it by volume. Projecting forward from requests per day rather than guessing a monthly total is the honest way to size the spend, because daily volume is the figure you can actually estimate from traffic and then scale to your growth plans.

The maths is transparent: per-request cost × requests per day, extended across an average month and a full year, so you can compare models and volumes on the number that ends up on the invoice.

What the projection assumes

The forecast assumes the volume you enter, growing by the monthly rate you set and otherwise steady, at standard list pricing. Real traffic is spiky, with weekday peaks, launch bumps and seasonal lulls, so treat the monthly figure as a midpoint and not as a limit. Where a vendor charges more for long prompts, the long-prompt rate applies when your input crosses the line. The projection leaves out what it cannot see from a list price: cached-input pricing for reused prompts, batch processing for non-urgent jobs, retries, tool calls and taxes. Use it to set a budget envelope and compare options, then confirm live rates with your provider. To compare the per-request cost across every model first, pair this with our LLM Price Comparison.

Disclosure: RECATOOLS is written with the assistance of Claude, made by Anthropic, whose models are priced on this page. Readers should weigh that interest.

An average month is 30.44 days, and this uses that

01

An average month is about 30.44 days (365.25 ÷ 12), so twelve of them make exactly the annual figure.

02

Output tokens usually cost several times more than input tokens on the same model, so reply length moves the bill quickly.

03

Cached-input and batch pricing are not included here, and they can cut a production bill substantially.

Frequently Asked Questions

  • It takes your per-request cost (input tokens × input rate + output tokens × output rate), multiplies by requests per day to get a daily cost, then by an average month of 365.25 ÷ 12 days. The annual figure uses 365.25 days.
  • It is the average length of a calendar month, 365.25 ÷ 12 = 30.4375, shown rounded. The monthly and annual figures both come from that year of 365.25 days, so twelve months equal one year exactly.
  • Measure a representative prompt with our Token Counter for the input, and estimate output from the length of replies you expect. If you've already shipped, your provider's usage dashboard shows actual average input and output tokens per request.
  • Unless you set a growth rate, yes. Real traffic has peaks and lulls, so treat the monthly figure as a midpoint and budget some headroom for spikes.
  • No, it uses standard list rates. Cached-input pricing for reused prompts and batch processing for non-urgent jobs can lower the real bill, often substantially. That does not make the projection a ceiling: retries, tool calls and traffic spikes can push the bill the other way.
  • The Cost Calculator works from requests per month; this projector works from requests per day and adds an annual forecast and a monthly request count. They share the same pricing, so use whichever input you can estimate more confidently.
  • Yes. Enter a monthly growth rate and the tool compounds your daily volume month by month, shows all twelve months with a running total, and puts the twelve-month sum in the annual box. A negative rate models a decline.
  • They're list rates from our shared model rate card. The dates of the oldest and newest row check are printed under the model dropdown, and each vendor's own caveat appears there too when it has one. Providers change pricing without notice, so confirm the live rate before committing a budget, especially here, where a small per-request error is multiplied by a year.
  • No. All calculation happens in your browser. The tool does not send or store your token counts and volumes.

Method & sources

How it computes

Cost per request = input tokens ÷ 1M × input rate + output tokens ÷ 1M × output rate, with the vendor's long-prompt rate where the input crosses its threshold. Daily spend = cost × requests per day; a year is 365.25 days and a month a twelfth of that. An optional monthly growth rate compounds from month 1. Rates come from the shared rate card, whose oldest and newest row checks are shown (LlmCostKernel.project).

What this tool implements

  • The projection is arithmetic; accuracy rests on the rates, so the page prints the dates of the oldest and newest row check. Prices are per-row provenance: each model carries its own vendor pricing URL and verified date on the shared rate card, database/data/llm-rate-cards.php.
  • ⚠️ Published list prices only. Enterprise agreements, committed-use discounts and promotional rates are not visible to us and are frequently lower.
  • Cached-input and batch rates are not modelled on this page, nor are retries, tool calls or tax, so the projection is neither a floor nor a ceiling.
  • Long-prompt tiers are applied per request by input size, and a request larger than the model's context window is flagged rather than silently priced.
  • ⚠️ LLM pricing moves faster than almost any other input on this site. A rate card is a snapshot with a date on it, and the date is shown so a stale figure is visible rather than silent.
  • Corrected 2026-11-02: month and year now agree (365.25 days; the page had 30.44 and 365, so twelve months were 365.28 days); the claim that the projection is a conservative ceiling, against the same page calling it a midpoint; the default model, pinned to a previous generation; the rate note dropped vendor caveats and labelled xAI's tier as above 200K when it starts at it; unpriced rows could read as $0.00.

Sources

Advertisement
Pre-footer · AD-W3 728 × 90