Reasoning Token Cost Calculator
See the hidden thinking tokens behind a reasoning model, and the true cost they add to every call.
Reasoning models like o3, o1 and DeepSeek-R1 do their thinking in tokens you never see in the response, and those tokens are billed at the output rate. That is why your reasoning model is so expensive: a short visible answer can hide thousands of reasoning tokens behind it. This o3 thinking tokens cost tool estimates that hidden work. Pick a model, enter your input and visible output tokens, choose a reasoning effort, and it shows the extended thinking cost you actually pay versus the naive figure you would guess from the visible answer alone.
Read the guide: Why Reasoning Models Cost MoreYour reasoning call
True cost per call
$0.0198
3.5× the naive estimate on o3-mini
- Naive (input + visible)
- $0.005720
- True (with reasoning)
- $0.0198
- Hidden reasoning tokens
- 3,200
- Extra cost
- $0.0141 (+246%)
Where the money goes
Reasoning (or "thinking") tokens are generated invisibly — you never see them in the response, but the provider bills them at the output rate. The hidden-token multipliers here (×1 / ×4 / ×10) are rough estimates of public model behaviour, not published constants; real reasoning use varies by prompt. Prices updated January 2026 — verify against the provider's pricing page before you budget.
How it works
- 1
Pick the reasoning model
Choose the model you call. Its input and output rates show underneath — reasoning tokens are billed at that output rate, so a pricey output model makes thinking expensive fast.
- 2
Enter tokens and effort
Set the input tokens and the visible output tokens you expect, then choose low, medium or high reasoning effort. Higher effort assumes more hidden thinking tokens per visible token.
- 3
Read the true cost
The headline is the real cost per call with reasoning tokens included. The stats compare it to the naive input-plus-visible estimate and show how many hidden tokens and how much extra you are paying.
Instant & 100% private — nothing is uploaded
Every calculation runs locally in your browser. The prompts, token counts and numbers you enter stay on your own device and are never sent to a server — nothing is stored, logged or shared.
Frequently asked questions
- What are reasoning tokens?
- They are the tokens a model generates while thinking through a problem before it writes the answer. You do not see them in the response, but they are counted and billed at the model's output rate, so a call can cost several times what the visible answer suggests.
- Why is my reasoning model so expensive?
- Because the hidden thinking dominates the bill. On a hard prompt a model may generate ten times more reasoning tokens than the visible reply, and every one is charged at the output rate. The short answer you get back is only a fraction of what you paid for.
- How do the low, medium and high multipliers work?
- They set how many hidden reasoning tokens the tool assumes per visible output token: roughly 1x on low, 4x on medium and 10x on high. These are rough estimates of public model behaviour, not published constants — actual reasoning use swings with the difficulty of each prompt.
- Is this the exact cost my bill will show?
- No. Real reasoning token counts vary per request and providers do not publish a fixed ratio, so treat this as a planning estimate. For the exact number, read the reasoning token count returned in your API response and price it against the provider's current rates.
Important
For planning and estimates only. Prices come from a published rate table dated on the page; providers change pricing without notice, and token counts here are approximations. Confirm against the provider’s own pricing before you budget or commit.
More tools
More from the Hivly network
Free sister tools on our other sites.
