1. Home
  2. Everyday math
  3. Rounding Calculator
Everyday math

Rounding Calculator

Round a finite number to a chosen number of decimal places.

Your numbers

Results update as you type. Nothing you enter is sent anywhere.

Result

12.35

Rounded = round(value × 10ᵈ) ÷ 10ᵈ. 0–6 decimal places. Half ties round toward positive infinity, including negative values; floating-point arithmetic applies.

Formula used

Rounded = round(value × 10ᵈ) ÷ 10ᵈ

Round a number to between 0 and 6 decimal places. This is handy for prices, measurements, lab results and grades, where you need a fixed number of decimals. The calculator shifts the decimal point as text rather than multiplying in binary, so awkward cases like 1.005 and 2.675 round the way you would expect on paper.

How to use the Rounding calculator

  1. Enter the value in Number. Negative numbers are accepted.
  2. Enter how many digits to keep after the decimal point in Decimal places, as a whole number from 0 to 6. Use 0 to round to the nearest whole number.
  3. Read the rounded value. Large results are shown with comma thousands separators, such as 1,235.

How the rounding calculation works

FormulaRounded = round(value × 10ᵈ) ÷ 10ᵈ

How ties are handled

When the part being dropped is exactly half, a rule is needed. This calculator rounds halves toward positive infinity: 2.5 becomes 3, and −2.5 becomes −2. That matches the behavior of JavaScript's Math.round and is the ‘round half up’ most people learn in school, applied consistently to negatives.

Other tools use other rules. The NIST guide to SI units recommends rounding an exact half so the last kept digit is even, so 2.5 rounds to 2 and 3.5 to 4; Python's built-in round() also rounds ties to the even choice. Excel's ROUND moves negative ties away from zero: Microsoft's documentation lists ROUND(−1.475, 2) as −1.48, where this calculator would give −1.47. If your answer differs from another tool only on a value ending in 5, the tie rule is the reason.

Why 2.675 can round to 2.67 elsewhere

Computers store 2.675 as the nearest binary fraction, which is slightly less than 2.675. A naive method that multiplies by 100 and rounds sees 267.49999… and returns 2.67. Python's documentation describes exactly this case. This calculator moves the decimal point by editing the number's written exponent, so 2.675 is treated as 267.5 and rounds to 2.68.

To round to tens, hundreds or thousands, which this tool does not do directly, divide first, round to 0 places, then multiply back. For 48,372 to the nearest thousand: 48.372 rounds to 48, so the answer is 48,000.

Worked example

Rounding 2.675 to two decimal places

Number: 2.675 · Decimal places: 2

  1. Keep two decimals: the kept part is 2.67 and the next digit is 5.
  2. The dropped portion is exactly half a hundredth, so the tie rule applies and rounds up.
  3. Shift the decimal: 2.675 → 267.5 → rounds to 268 → shift back to 2.68.

Result: 2.68

Assumptions and limits

0–6 decimal places. Half ties round toward positive infinity, including negative values; floating-point arithmetic applies.

Formula and worked examples last checked November 20, 2026. Spotted a problem? Report an error.

Frequently asked questions

What is 3.14159 rounded to two decimal places?

It is 3.14. The third decimal digit is 1, which is less than 5, so the second decimal stays as it is. Rounded to four places it would be 3.1416, because the fifth digit is 9.

Does 0.5 round up or down?

In this calculator 0.5 rounds up to 1 and −0.5 rounds to 0. Under round-half-to-even, used by NIST's rounding guidance and by Python, 0.5 rounds to 0 and 1.5 to 2. Check which rule your teacher, lab or software expects.

How do I round to the nearest whole number?

Set Decimal places to 0. Then 17.49 becomes 17 and 17.5 becomes 18. Only the first digit after the decimal point decides the outcome.

Should I round in the middle of a calculation?

Avoid it. Rounding each step can add up to a noticeable error; for example, rounding 1/3 to 0.33 and multiplying by 300 gives 99 instead of 100. Keep full precision until the final answer, then round once.

Sources