CURBTOLL

TOOL / COMPARE

ATM vs Exchange Calculator

Enter an amount and the fee tiers, and see how much bitcoin each route leaves you with. The market price comes from a public feed; the fee figures are yours to change to whatever the machine in front of you shows.

Default figures checked 2026-09 Everything runs in your browser

Calculator cover: bars representing the two routes and the gap between them
Put your own number in.

The same cash gets you different amounts of bitcoin depending on the route. How different comes down to four numbers.

Machines don't give change — whatever goes in gets converted.

Reported range as of 2026-09: 10–25, centre near 16.

Never shown on screen. Commonly 3–5.

Mainstream spot fees are usually under 1%. Check your platform's fee page.

Loading… Binance public price feed

The feed gives a US dollar price, so the amount above is read as US dollars. Any currency works — set the market price to that currency's price and enter your amount in it, and the percentages and gap all still hold.

Via the machine
Via an exchange
Bitcoin you don't get
In money
WAITING ON YOU

Put in an amount and you'll see the gap.

The market price is fetched from Binance's public endpoint; if the request fails the tool switches to manual entry rather than using any stored figure. Fee percentages are entirely yours — the tool makes no assumption about what any machine charges. All calculation is local to your browser.

Which line to look at

The one marked "in money". Percentages go numb quickly; an amount doesn't.

  • The defaults are a starting point. Standing at a machine, put its actual posted fee in. The spread can't be read off the screen — leave 4% as an estimate, or work out the real figure with the spread checker.
  • This compares coins received, not profit. Both routes face identical price movement afterwards. The difference is where you start — ten to twenty per cent fewer coins means a smaller position from day one, whichever way the market goes.
  • The network fee isn't included. It's a fixed amount and hurts small purchases hardest, so the real gap is usually a little wider than shown.

How big a gap should change your mind?

No universal answer — it depends how much you need what the machine offers: speed, cash acceptance, no account.

A useful framing: treat the "in money" figure as the price of a ticket for "today, and without a bank account". Then ask whether you need that ticket.

Bank account and no deadline? You bought a ticket and didn't board. No other route available? Then it's money spent on something real — the short list sets out when that's the case.

Common questions

Is the market price in this tool real?

Yes. It is fetched from Binance's public price endpoint when the page loads. If the request cannot be completed the tool switches to a manual input box for you to enter a price yourself, and it never falls back to a stored or hard-coded value.

Where do the default fee figures come from?

They are mid-range values from publicly reported industry data as of September 2026: a 12% posted fee, a 4% rate spread and a 0.1% exchange fee. They are only a starting point and do not represent any particular machine, so replace them with whatever the screen in front of you shows.

Why is the gap larger than the machine's stated fee?

Because the machine's cost has two layers. The posted fee is disclosed, but the exchange rate it quotes carries a further margin that is not itemised anywhere. Both are included in this calculation, so the total is higher than the percentage displayed on the machine.

Can I use this with Australian dollars?

Yes. Switch the market price to manual entry, put in the price per bitcoin in Australian dollars, and enter your amount in Australian dollars too. As long as both figures use the same currency, the percentages and the money gap are all correct.

Default ATM inputs refer to Coincub's August 2026 analysis, and the exchange input can be checked against Binance's current fee table. They are starting points only; change them to match the screen. Nothing is uploaded.