AI Model Comparison

Interactive tool

LLM Cost Calculator: Price Your Token Mix on 9 Models

Free LLM cost calculator: compare the monthly bill on nine models at once with your own token mix, cached-input rates included. Checked 22 November 2026.

Standard API rates · USD per 1M tokens · checked 2026-11-22
ModelInput $/1MOutput $/1MCost of your workload

Standard rates, USD per million tokens, checked 22 November 2026. Cached-input rates where the provider publishes one. Preview promotions are marked.

Pick a model, put in your real token counts, and see what the month costs. The calculator runs on the standard rates from our comparison table — the same figures, checked 22 November 2026, with the cached-input rate applied where the provider publishes one. Nothing leaves your browser. If you came looking for an llm comparison tool rather than another price table, this is it: nine models, your numbers, no signup.

Three things the number above is telling you:

  1. Output tokens dominate most bills. Output runs 3–12× the input rate across the market. A workload that writes 50K tokens costs more than one that reads 500K on most models here.
  2. Cached input is the biggest lever for agents. If your system prompt is the same on every call, move the cached share slider up. On Inkling 2 Mini, cached input is $0.015 per million — 100× cheaper than an uncached frontier input.
  3. Per-request price is not per-month price. A model that looks expensive at $40 output can be cheap at 1,000 requests a month if it writes few tokens; a cheap model that retries twice is not cheap.

How the calculation works

Cost per request = (fresh input tokens × input rate + cached input tokens × cached rate + output tokens × output rate) ÷ 1,000,000. Monthly cost = per-request cost × requests. Batch rates (roughly half price, slower returns) and free tiers are not included — the table shows standard rates only.

When the calculator's answer is wrong

When you have not measured your real token counts. Token counts vary by tokenizer: the same document is a different number of tokens on different models. Use your provider's own usage dashboard numbers, not estimates, when you feed this page. And budget retries: if you expect to escalate failures to a frontier model, add that share by hand — a cheap-first setup is usually cheaper than it looks here.

Where to go next

Frequently asked questions

How much does an LLM API cost per month?

Your monthly tokens × the per-million rate. One million output tokens costs between $0.50 (Inkling 2 Mini) and $40.00 (GPT-6.2 Astra, Claude Fable 5.2) across this index — an 80× spread, which is why the calculator exists.

What is a token worth in dollars?

There is no single answer: input tokens, output tokens and cached input tokens are billed at different rates on every model. The calculator applies all three rates to your token mix.

Is prompt caching worth it?

When the same context repeats across calls, yes: cached input is 10–100× cheaper than fresh input on the models that support it. It changes nothing for one-off requests.

Does the calculator include batch or free tiers?

No. Batch rates return results later for roughly half price; free tiers are rate-limited. Both are real savings, but they are not standard rates, so they stay out of the comparison.