Rounding calculator

Choose the kind of precision you need

Decimal places count digits to the right of the decimal point, so 12.345 rounded to two decimal places is 12.35. Integer place rounds to ones, tens, hundreds, or a larger power of ten; entering 2 means the nearest 10², or hundred. Significant figures count meaningful digits from the first nonzero digit, so leading zeros in 0.004567 do not consume the selected precision.

Type the number, choose a mode, and enter its precision. Press Calculate or Enter to confirm the current values. Editing the number, mode, or precision clears the old answer and disables Copy result, which prevents a previous result from looking current. Clear removes the number and result while keeping your selected mode and precision.

1. Enter a plain decimal or E-notation value.
2. Choose decimal places, integer place, or significant figures.
3. Set the requested precision and select Calculate.
4. Read the result and rule explanation, then copy it if needed.

The rounding rule used here

The tool rounds to the nearest representable value at the selected place. It examines the first discarded digit once: 0–4 keeps the retained magnitude closer to zero, while 5–9 increases the retained magnitude. Exact halfway values use round half away from zero. Therefore 2.5 becomes 3 and -2.5 becomes -3 when rounding to zero decimal places. The negative result does not use JavaScript’s asymmetric Math.round behavior.

Input is read as base-ten text and rounded with integer arithmetic. That matters for values such as 1.005: a naive binary floating-point calculation may expose an approximation just below the halfway point, but this calculator rounds the decimal you typed and returns 1.01 at two decimal places. It rounds once at the requested boundary and does not repeatedly round intermediate digits.

Decimal places, integer places, and significant figures

Decimal-place output keeps the requested trailing zeros because they communicate precision: 12 rounded to three decimal places is displayed as 12.000. Integer-place output replaces discarded positions with zeros. Select 0 for the nearest integer, 1 for the nearest ten, 2 for the nearest hundred, and so on. Carrying is handled across digits, so 9.999 to two decimal places becomes 10.00 and 9,950 to the nearest hundred becomes 10,000.

Significant figures are useful when the magnitude varies. Rounding 0.0004567 to two significant figures produces 0.00046, while 45,670 to two significant figures produces 46,000. Zero has no first nonzero digit, so the selected precision cannot be inferred from magnitude alone; this page displays zero with explicit decimal zeros where possible, such as 0.00 for three significant figures. For ambiguous large whole numbers, record the selected precision with the copied context or use scientific notation in formal work.

Input limits and responsible use

The field accepts a sign, digits, one decimal point, and an optional E exponent from -300 to 300. A coefficient may contain up to 200 digits. Commas are rejected because they can mean either a thousands separator or a decimal mark in different locales. Remove currency symbols, percent signs, units, spaces inside the number, and arithmetic expressions before calculating. All six language versions use a period as the decimal separator so the same input is interpreted consistently.

Rounding changes presentation and may discard information; it does not improve the accuracy of a measurement. Keep the unrounded source for later calculations and round the final reported value unless your method requires otherwise. This page supports one common nearest-value rule only. It does not provide ceiling, floor, truncation, half-even banking rules, uncertainty propagation, or a general expression parser. Financial, scientific, tax, and regulatory work may specify a different rule.

Using and sharing the result

The result card shows the entered decimal value, the selected precision, the first discarded digit when one exists, and the decision made by the rule. Copy result places the rounded value and its precision context on the clipboard instead of copying a bare number whose trailing zeros could be misunderstood. The calculation runs locally in the browser, no history is saved, and changing language may reload the page and clear the form.

For a spreadsheet or report, confirm whether the receiving system preserves trailing zeros and applies its own display formatting. Store the exact source separately from the rounded display. A displayed zero may mean the input rounded to zero at the chosen place, not that the original number was exactly zero.