Written by Sarah Mitchell · Reviewed by Alex Chen · July 31, 2026
Count the number of significant figures in any value and round to a specified precision.
Enter a number and click Count.
Enter a number and sig figs count, then click Round.
Significant figures are the digits of a number that carry information about its precision. The leading zeros that only locate the decimal point do not count; the trailing zeros that were actually measured do. This page counts significant figures in any number you enter and can also round a value to a chosen number of significant figures, giving you both the reading and the rewriting in one step.
Type a number into the top field and the count appears immediately next to it. The tool looks at every digit, skips the zeros that only set the scale, and reports how many digits genuinely represent the measurement. To round instead, enter a number in the lower field, choose how many significant figures you want to keep, and read the rounded value below. The round mode works with any positive count of significant figures.
Enter 1.230e4 and the page reports 4 significant figures, because the zero after the three is part of the measured value and the scientific notation simply moves the decimal point. Enter 3.10 and the count is 3: both digits and the trailing zero, which says the measurement was taken to the hundredths place. Enter 1.5e-3 and the count drops to 2, since the two zeros only position the decimal point and carry no information about precision.
Plain numbers follow the same rules. The value 0.0045 has 2 significant figures, and 12300 has 5. The tool counts the digits that mean something and ignores the zeros that only set the size.
Zeros behave differently depending on where they sit. A zero between two nonzero digits always counts, so 105 has three significant figures. A zero after the decimal point and after other digits counts, which is why 3.10 has three. A zero before the first nonzero digit never counts, because it only indicates the place value; 0.0045 loses both of its leading zeros. The page applies these conventions consistently, and the count updates on every keystroke so you can watch the rules in action.
Rounding to significant figures follows the same logic as rounding to a decimal place, but the target moves. The tool works out where the decimal point sits, keeps the requested count of meaningful digits, and rounds the rest using the usual rules. Rounding 123.45 to 3 significant figures gives 123, rounding 0.004567 to 2 gives 0.0046, and rounding 2.345 to 3 gives 2.35. Negative numbers round with their sign intact, so -3.10 rounded to 3 significant figures stays -3.1.
Many textbooks teach that a number like 12300 has 2 or 3 significant figures because the trailing zeros might just be placeholders. This page follows the opposite convention: it counts every digit in the input as meaningful, so 12300 reports 5. If you want to be explicit about precision, scientific notation is the clearer route, and the page's count for 1.23e4 is 3. The rule of thumb is simple: when trailing zeros matter, write the number in scientific notation, and the counter will agree with your intent.
Because they do not measure anything; they only place the decimal point. The value 0.0045 has the same measured digits as 4.5, so both carry two significant figures.
It makes precision explicit. The digits in the coefficient are all significant, so 1.230e4 has four and 1.2e4 has two. Entering scientific notation is the clearest way to tell the page how many figures you mean.
It keeps the first two meaningful digits and rounds the rest, giving 0.0046. The leading zeros survive untouched because they only set the size of the number.
The minus sign is preserved through both counting and rounding. The significant figure rules apply to the digits themselves, and the result keeps the sign it started with.
Yes. The lower field rounds on its own, independent of the counter above it, so you can go straight to a rounded result even if you did not check the count.
Significant figures exist because measurements have limits. A scale that reads to the tenth of a gram cannot truthfully report a hundredth, and a ruler measured in millimeters cannot support a result claimed to the micron. Science labs, engineering reports, and recipe conversions all rely on stating only the precision you actually have. This page makes that check instant, counting and rounding any number in the browser with no data sent to a server.
Laboratory glassware advertises its own precision. A graduated cylinder marks every milliliter, a burette reads to a hundredth, and a digital balance displays four decimal places, yet none of them reports more truth than the instrument can sense. If a thermometer shows 21.4 degrees, claiming 21.43 from the same glance overstates the measurement, because the last digit was not actually observed.
That gap between what a tool shows and what a result may claim is where significant figures do their work. Analysts record every digit they can defend, then let the count drop when they multiply, divide, or combine readings, so the final answer never pretends to more certainty than the weakest measurement supplied.
Uncertainty is the quiet partner of every measurement. A balance that drifts by a few milligrams, a stopwatch that lags by a fraction of a second, or a ruler with a chipped end all inject error before any arithmetic begins. Good record keeping notes the margin of error alongside the value, and the significant figure count is the shorthand that carries that caution into the final result.
The same logic runs through ordinary objects. A kitchen scale that steps in grams cannot distinguish 1.4 grams from 1.5 grams, a GPS unit that rounds distance to tenths of a mile hides any finer motion, and a thermometer that marks whole degrees leaves fractions to guesswork. When a device truncates, the extra digits it prints are decoration rather than data.
Currency behaves differently because money is exact by decree: a price in dollars and cents has every digit defined, so 10.50 carries four significant figures, not three. Distinguishing measured values from counted ones is the heart of the convention, and running a number through the page's counter makes that distinction visible at a glance.