Glossary
What is API-equivalent cost?
API-equivalent cost is what your AI coding tokens would cost at the provider's published API rates. What it measures, how it is computed, and what it is not.
· Facts checked October 3, 2026
The short answer
API-equivalent cost is the price of your token usage at the provider's published API rates, whatever plan you actually pay for. It is an estimate for comparison, not a charge: on a Claude or ChatGPT subscription you pay the plan price, and the API-equivalent figure shows how much usage that plan delivered.
At a glance
| Formula | Each token category times its published rate: input, cache reads, cache writes, and output. |
|---|---|
| Useful for | Comparing plans, models, and tools on one scale. |
| Not the same as | A subscription charge, an invoice, or spend a provider reports. |
| Changes when | Providers change their published rates. |
How it is calculated
Every request a coding tool makes is split into token categories, and each is priced separately. With Claude Sonnet 5, for example, a million ordinary input tokens cost $2, a million cache reads $0.20, a million 5-minute cache writes $2.50, and a million output tokens $10. Adding the categories gives the request's API-equivalent cost; summing requests gives a day or a month.
Pricing cache writes and reads separately matters, because a figure that prices every input token at the ordinary rate overstates the cost of a cached session by a wide margin.
What it is good for, and what it is not
On a subscription, the number answers a value question: a month of Claude Code worth $300 at API rates on a $100 plan returned three times its price, which is the core of the Max plan guide. It also puts tools with different billing models on one scale.
It is not what you paid. It ignores taxes, batch and priority tiers, discounts, and the plan price itself. Tools that report their own spend, such as OpenRouter for a key or Cursor for its plan cycle, should be shown as reported and kept apart from estimates, which is how everyaitoken displays them.
Your own numbers
See what your AI coding really costs.
everyaitoken reads Claude Code, Codex, Cursor, OpenRouter, Gemini CLI, and OpenCode on your Mac. It shows limits with reset times, API-equivalent cost, and what caching saved or cost, in the menu bar and 17 widgets. $9 once.
FAQ
Questions
Is API-equivalent cost what I pay?
No. It is what the same tokens would cost at API rates. On a subscription you pay the plan price; on the API you pay the provider's actual charges, which can include taxes and tier adjustments.
Why does Claude Code show a cost on a Max plan?
The session cost Claude Code shows is meant for API users. Anthropic notes it is not relevant to subscription billing, but as an API-equivalent figure it shows how much usage the plan delivered. See Claude Code cost per day.
Can API-equivalent cost change after the fact?
Yes, if it is recomputed with updated rates. A provider price change applied to past tokens changes the historical estimate, which is why the rates used should be dated.
Sources
Checked October 3, 2026. Plans and limits change; the linked pages are the authority.