buildagentic

Calculator

What will this model cost me a month?

Per-token prices are quoted per million tokens, which makes two models an awkward comparison and a monthly bill an awkward guess. This turns a volume and a split into a figure, from the prices we sync from OpenRouter.

How this is worked out

  1. 1

    Take the model's input and output prices per million tokens, synced from OpenRouter.

  2. 2

    Split your monthly volume by the output share you set.

    output = volume x share input = volume x (1 - share)
  3. 3

    Charge each half at its own rate and add them.

    monthly = input x price_in + output x price_out
Every figure here comes from Language and embedding models — 18 tracked, each with its source and date.last sync 22h ago →

Your numbers

Everything updates as you change it, and the address bar keeps up — so the answer you are looking at is the answer you can send someone.

100M tokens

Everything in and out. A small internal tool is around 20M; a customer-facing bot, hundreds.

25%

Replies cost more than prompts. A chat assistant is around 25%; a summariser closer to 5.

Model cost per month

$1,000per month

75M input at $5.00 and 25M output at $25.00, per million tokens.

Synced from sourceInput price
$5.00per MtokCurrent22h ago
Synced from sourceOutput price
$25.00per MtokCurrent22h ago
Synced from sourceContext window
1,000,000tokensCurrent22h ago
Volume
100 Mtokens per month
Output share
25%
All 3 inputs synced from a source.
Monthly cost
Claude Opus 5

Line height = what you would pay that month at that volume.

$0.00$500.00$1,000$1,500$2,0000M50M100M150M200MTokens per monthMonthly costYou: 100MMonthly costMonthly cost$0.00Tokens per month: 0MAll 3 inputs synced from a source.Monthly cost$166.67Tokens per month: 17MAll 3 inputs synced from a source.Monthly cost$333.33Tokens per month: 33MAll 3 inputs synced from a source.Monthly cost$500.00Tokens per month: 50MAll 3 inputs synced from a source.Monthly cost$666.67Tokens per month: 67MAll 3 inputs synced from a source.Monthly cost$833.33Tokens per month: 83MAll 3 inputs synced from a source.Monthly cost$1,000Tokens per month: 100MAll 3 inputs synced from a source.Monthly cost$1,167Tokens per month: 117MAll 3 inputs synced from a source.Monthly cost$1,333Tokens per month: 133MAll 3 inputs synced from a source.Monthly cost$1,500Tokens per month: 150MAll 3 inputs synced from a source.Monthly cost$1,667Tokens per month: 167MAll 3 inputs synced from a source.Monthly cost$1,833Tokens per month: 183MAll 3 inputs synced from a source.Monthly cost$2,000Tokens per month: 200MAll 3 inputs synced from a source.
The models dataset →
Input
$375.0075M tokens
Output
$625.0025M tokens
Per day
$32.89

Background

Why per-million pricing is hard to budget from

Every gateway quotes dollars per million tokens, which is a fine unit for comparing two models and a poor one for answering "what will this cost me". A million tokens is not an amount anybody has an intuition for, and the number you actually want — a monthly line item — depends on a split nobody quotes: how much of your traffic is the model reading versus the model writing.

That split matters more than it looks. Output is typically the dearer half, and the gap between a summarising workload at 5% output and a conversational one at 40% can be most of the bill. Two teams on the same model at the same volume can pay very different amounts, which is why a single per-token headline is not a budget.

The prices here are synced from OpenRouter and carry the moment they were read. Gateway pricing moves, models are withdrawn, and a figure quoted from a cached page is worse than no figure — so if we have no current price for a model, this tool says so rather than estimating one.

What this tool assumes

Computes
Monthly spend at your volume and split
Prices
Live, synced from OpenRouter
Split
Input and output charged separately
Excludes
Cached-token discounts, batch tiers
Refuses
Any model we hold no current price for

The data underneath

Language and embedding models

Every number this calculator used is a row you can open, with the source that reported it and the moment it was read. Nothing here is an estimate we could not show you.

Open the dataset →

More tools

Other calculators

Questions

What people ask

What counts as a token?

Roughly three quarters of a word in English, for both what you send and what comes back. A page of text is about 500 tokens; a long chat exchange, a few thousand.

Why is my real bill different?

Most likely cached input, batch pricing, or a different split than you set here. This computes list price at the split you gave — it is a budgeting figure, not an invoice.

How current are these prices?

The freshness badge beside each price says exactly, in hours. They sync from OpenRouter, and a model whose price we cannot currently read is refused rather than estimated.

Can I compare two models?

Change the model and the URL changes with it, so two tabs is the fastest way. The models dataset also has a side-by-side comparator for the underlying figures.