Significant Figures Calculator
Count the significant figures in a number as you wrote it, or round it to a count you choose. The notation is the input that matters — 1500, 1500. and 1.500 × 10³ are one value making three different claims — so the tool reads your text, not a parsed number.
Count or round to sig figs
Both answers are worked out either way — this picks which one leads.
Type it exactly as written. 1500, 1500. and 1.500e3 are the same value but three different claims about precision.
How many figures the rounded answer should carry. Asking for more than the number has pads it with zeros — a claim, not a fact.
Every digit from the first non-zero one onward counts, which makes 0.004560 the same precision as 4.560 × 10⁻³.
- Significant figures
- 4
- Decimal places
- 6
- Scientific notation
- 4.560 × 10⁻³
- Rounded
- 0.00456
- Rounded, scientific
- 4.56 × 10⁻³
Three rules that always work, and one case that has no answer
Significant figures exist for one reason: to stop a calculation claiming more precision than its inputs had. A tape measure reading 2.5 m multiplied by 3.14159 gives 7.853975 on any calculator, and that answer asserts a measurement to the nearest micrometre from a tool that could not resolve a centimetre. The rules below are how a number is made to carry its own honesty.
0.00456 → 3 • 10.05 → 4 • 0.004560 → 4 • 1500 → 2 to 4- Leading zeros
- Never significant — they only place the decimal point
- Enclosed zeros
- Always significant — a zero between two non-zero digits was measured
- Trailing zeros, decimal present
- Always significant — there is no other reason to write them
- Trailing zeros, no decimal
- Ambiguous — the notation genuinely cannot say
Why the count is taken from your text, not from the value
Significance is a property of the notation, not of the quantity. 1500 and 1500. are the same number and different claims. Parse either one into a double and both arrive as the identical sequence of bits, with no trace of which was typed — so this calculator counts the characters first and converts afterwards. That is also how scientific-notation input works properly here: 1.5 × 10³ is two figures and 1.500 × 10³ is four, and a tool that parsed first would have to call them both the same.
Why rounding is done on the digits
The rule taught in every science class is half-away-from-zero: if the part being dropped starts with a 5, the last kept digit goes up in magnitude. The usual implementation — multiply by a power of ten, call the language rounding function, divide back — fails on exactly the cases where the rule is being invoked, because most decimal fractions have no exact binary form. The value 1.005 is stored as 1.00499999999999989, so that approach returns 1.00 at three significant figures. Rounding the digit string with a longhand carry instead is exact: 1.005 gives 1.01, and −2.5 at one figure gives −3 rather than −2. It also lets the answer keep its trailing zeros, which a number cannot — 2.5 at four significant figures is 2.500, and there is no floating-point value that holds that claim.
2.5 m of shelf, and a circle of it
A tape measure reads 2.5 m. That is two significant figures, and it means the true length is somewhere between 2.45 and 2.55 m — the last digit is the uncertain one, which is what “significant” is describing.
Multiply by π: 2.5 × 3.14159 = 7.853975. Every digit after the first two is manufactured. The multiplication rule says the result keeps the significant figures of the least precise input, and 2.5 has two, so the answer is 7.9 m. Check it against the interval: the honest product runs from 2.45 × 3.14159 = 7.70 to 2.55 × 3.14159 = 8.01, and 7.9 sits in the middle of that spread while 7.853975 pretends the spread does not exist.
Now the other rule. Add a 0.3 m offcut to a 12.11 m board and the arithmetic gives 12.41. Addition keeps decimal places rather than significant figures, and 0.3 has one decimal place, so the answer is 12.4 m — four significant figures, more than either input has on its own, which is correct and surprises people. Lining numbers up at the decimal point is a different operation from lining them up by relative precision.
Try the default in the tool above: 0.004560 counts as four significant figures, restates as 4.560 × 10⁻³, and rounds to 0.00456 at three. Switch it to 1500 and the count becomes a range, because at that point the notation has stopped saying.
What this calculator assumes
- Every number you type is a measurement. Exact counts and defined constants have unlimited significance and should be left out of the rules; the tool has no way to tell them apart from measured values.
- Halves round away from zero, so 2.5 at one significant figure is 3. ASTM E29 and the IEEE 754 default both round a bare half to the nearest even digit and would give 2. Neither is wrong — they answer to different traditions, and this one matches what is taught in school science.
- Commas and spaces are stripped as grouping marks before counting, so 1,500 reads exactly as 1500. They never make a zero significant.
- A trailing decimal point is honoured as deliberate: 1500. means four significant figures. It is a real convention but not a universal one, so prefer scientific notation when the reader is not expecting it.
- Zero is reported as one significant figure. The rules are built around the first non-zero digit, and zero has none, so any answer here is a convention rather than a consequence.
- Nothing propagates through a calculation. This page counts and rounds one number at a time; applying the addition and multiplication rules needs the whole expression, and that is yours to do.
Significant figures FAQ
How many significant figures does 0.004560 have?
Four. The three zeros at the front are placeholders — they only say where the decimal point sits, and 0.004560 is the same precision as 4.560 × 10⁻³, which plainly has four digits. The zero at the end is different: it is after the decimal point and after a non-zero digit, so the only reason to write it is that it was measured. Leading zeros never count, trailing zeros after a decimal point always do.
Is 1500 two significant figures or four?
The notation cannot tell you, and any calculator that answers confidently is guessing. It depends on whether the value was measured to the nearest hundred or to the nearest unit. This is the one genuinely ambiguous case in the rules, which is why this tool reports a range — two through four — rather than picking. Two notations resolve it: write 1500. with a trailing decimal point to mean all four digits count, or write 1.5 × 10³ for two and 1.500 × 10³ for four. Resolving exactly this is a large part of why scientific notation exists.
Do I round 2.5 to 2 or to 3 at one significant figure?
To 3, under the rule taught in school science: when the part being dropped is exactly a half, round away from zero. So 2.5 becomes 3 and −2.5 becomes −3. Note that this is not the only convention. ASTM E29 and the default mode of IEEE 754 both round a bare half to the nearest even digit, which would give 2 — a rule that avoids the slight upward bias you get from always rounding halves up across a large data set. This calculator uses half-away-from-zero because that is what is being asked for in almost every case a person types a number into a web page.
What are the significant figure rules for adding and multiplying?
They are different rules, and mixing them up is the most common mistake. Addition and subtraction keep decimal places: 12.11 + 0.3 = 12.41 becomes 12.4, because 0.3 has only one decimal place. Multiplication and division keep significant figures: 2.5 × 3.14159 = 7.853975 becomes 7.9, because 2.5 has only two significant figures. The intuition is that adding lines the numbers up at the decimal point, while multiplying lines them up by relative precision.
Why does my calculator show 7.853975 when I measured 2.5 metres?
Because the calculator does not know that 2.5 was measured. It sees a number with an exact binary value and multiplies it by another. But 2.5 m means somewhere between 2.45 and 2.55 m, so the honest product is somewhere between 7.70 and 8.01 — and writing 7.853975 claims a precision to the nearest micrometre that the tape measure never had. Significant figures exist to stop that claim being made by accident. The answer is 7.9 m.
Do exact numbers have significant figures?
No — counts and defined constants have unlimited significance and never limit a result. If you counted 12 bolts, that 12 is exact, not 12 ± 0.5. The same goes for defined conversions such as exactly 2.54 centimetres to the inch. This tool has no way to know which of your numbers are counts, so it treats everything you type as a measurement. Exclude the exact values yourself when you apply the arithmetic rules.
Sources and review notes
- NIST Special Publication 811 — Guide for the Use of the International System of Units, §7.9 on rounding and the number of digits a reported value should carry
- ASTM E29 — Standard Practice for Using Significant Digits in Test Data, the source for the round-half-to-even convention noted above
- BIPM SI Brochure — the international definitions behind measured quantities and their stated uncertainty
The rules on this page are conventions rather than data, so nothing here expires. What can go wrong is the implementation, so the engine is a pure function under unit test — including the 1.005 case, which fails on any version that rounds through a floating-point multiply.