How Age Calculators Handle Leap Years.

01—Foundations

What “exact age” actually means

Ask someone their age and they’ll give you a whole number: “I’m 26.” But that whole number is really a rounded-down version of a more precise answer: 26 years, some number of months, and some number of days. That fuller breakdown is what an exact age calculator is solving for, and it’s a genuinely different calculation than the simple year subtraction most people do in their head, one that also has to get a leap year birthday right rather than fudging it.

The reason it trips people up isn’t the concept. Everyone understands “years, months, and days since I was born.” It’s the arithmetic underneath, because calendar months don’t all have the same length, and that irregularity has to be handled correctly every single time, not just in the easy cases.

02—The easy half

The straightforward part: subtracting years

Start with the year subtraction. If someone was born in 2000 and you’re calculating their age as of 2026, the starting point is 2026 − 2000 = 26 years. This part rarely causes problems; it’s the same subtraction anyone would do by hand.

The complication shows up in the next step, when you try to account for the months and days left over after that whole-year subtraction.

03—Borrowing

Where it gets hard: borrowing a month

Here’s the situation that breaks naive age math: what happens when the target date’s day-of-month is earlier than the birth date’s day-of-month? Try calculating the age of someone born January 31, 2000, as of March 5, 2026.

A naive approach subtracts day 31 from day 5 and gets negative 26 days, which isn’t a valid answer. The fix is the same trick used in ordinary column subtraction when the top number is smaller than the bottom one: borrow from the next column over. Here, that means borrowing a month, using that specific month’s real day count.

Working through it: borrow from February (28 days in 2026, since 2026 isn’t a leap year), giving 28 + 5 − 31 = 2 days, and reducing the month count by one. Combined with the year subtraction, the final answer is 26 years, 1 month, 2 days. Getting this right requires knowing exactly how many days were in the borrowed month, not a flat 30-day assumption, which is exactly where a lot of simplified age calculators go wrong. A 30-day-average approach would produce a noticeably different, and wrong, day count for any calculation that borrows from February, or from any 31-day month.

04—Try it

Try it yourself

Enter any birth date below and see the exact breakdown calculated live, using the same calendar-accurate method described above. This is the actual calculator, not a mockup.

05—Leap years

Leap years: the other edge case

The second place age math commonly breaks is February 29 birthdays. In a leap year, February has 29 days, so a February 29 birthday lands exactly where you’d expect. But three years out of every four, February only has 28 days, so there’s no February 29 to land on at all.

This calculator doesn’t hard-code either rule. Because the month-borrowing logic described above treats February 29 the same way it treats any other day-of-month comparison, the outcome falls out of the arithmetic rather than a special case: comparing day 29 against day 28 always finds the target day “earlier,” triggers a borrow, and that borrow only resolves on March 1. So the birthday isn’t skipped, but it also isn’t anchored to February 28 as some competing calculators do; it lands on March 1, which happens to match the convention several legal systems use for the same reason (see below).

06—Disagreement

Why two age calculators can disagree by a day

If you’ve ever compared two different age calculators and gotten answers a day apart, it’s almost always one of these two situations: a month-borrow calculation near a month-end birth date, or a February 29 birthday being handled differently by each tool. A calculator using a simplified 30-day-month approximation will drift further from the calendar-accurate answer the further apart the two dates are, since every borrowed month compounds the error by whatever the real month length differs from 30.

The safest check, if you want to verify an age calculation by hand: work through the year subtraction first, then check whether the day-of-month comparison requires borrowing, and if it does, look up the actual number of days in the specific month you’re borrowing from: not an average, the real number for that month and year.

07—The rule itself

The Gregorian rule, and the century trap

Almost everyone knows that a leap year happens every four years. Far fewer know that the rule has two exceptions stacked on top of it, and the second exception is the one that quietly breaks home-made date code.

1. 01

   Divisible by 4 → leap year

   2024, 2028 and 2032 all qualify on this rule alone. This is where most people stop, and for any date range inside a single human lifetime it is usually enough.

2. 02

   Unless divisible by 100 → not a leap year

   1900 was divisible by 4 but was not a leap year. Neither will 2100 be. This exists because the year is not exactly 365.25 days long. It is about 11 minutes shorter, and those minutes accumulate.

3. 03

   Unless divisible by 400 → leap year after all

   2000 was divisible by 100, so rule two would have excluded it, but it is also divisible by 400, so it was a leap year. This third rule corrects the slight over-correction of the second.

The reason for the layering is astronomical rather than arbitrary. A tropical year runs roughly 365.2422 days. Adding one day every four years assumes 365.25, which overshoots by about 11 minutes a year, or three days every four centuries. Skipping three leap days per four hundred years brings the calendar back in line to within about one day in 3,000 years.

Century years and whether they are leap years

Year	÷ 4	÷ 100	÷ 400	Leap year?
1900	yes	yes	no	No
1996	yes	no	no	Yes
2000	yes	yes	yes	Yes
2024	yes	no	no	Yes
2100	yes	yes	no	No
2400	yes	yes	yes	Yes

08—Leap-day births

What happens if you were born on 29 February

Roughly one person in 1,461 is born on a leap day, and for three years out of every four their birthday does not exist on the calendar. This is a genuine ambiguity rather than a bug, and different systems resolve it differently, legitimately so, because there is no mathematically correct answer.

Neither convention is wrong; they answer slightly different questions. The first asks “which day best corresponds to 29 February?” and the second asks “on which day has a complete year definitely passed?” Where the answer matters legally (drinking age, retirement eligibility, contract terms), the governing jurisdiction specifies which applies, and it is not consistent internationally. England and Wales settled it by statute: Section 9 of the Family Law Reform Act 1969 fixes the moment someone attains a given age as the start of the relevant anniversary of their birth, which for a 29 February birth in a common year is 1 March, not 28 February. (The provision took effect from 1 January 1970 and doesn’t extend to Scotland.) That’s the second convention above, and it’s the one this site’s calculator produces as well, for the arithmetic reason explained earlier rather than because it was deliberately coded to match UK law.

1 in 1,461

For an age calculator the honest behaviour is to pick one convention, state it, and apply it consistently. Silently switching between them, which happens when code borrows days in one branch and months in another, produces a tool that reports different ages for the same person depending on what today’s date happens to be.

09—A concrete example

How old is a leap-day baby right now, and how many real birthdays have they actually had

Two different numbers answer two different questions here, and people asking about leap-year age usually want the first one without realizing there’s a second.

The first is ordinary calendar age: someone born 29 February 2000 is 26 in 2026, counted the same way as anyone else born in 2000, using whichever March-1-or-February-28 convention the calculator applies in the years February 29 doesn’t exist. The second is how many times February 29 has actually appeared on the calendar since they were born, which is the number people mean when they joke that a leap-day baby is “only 6” on their actual birthday.

Leap-day (29 February) births and where they stand in 2026

Born 29 Feb of	Real age in 2026	Actual Feb 29 birthdays so far
2024	2	0
2020	6	1
2016	10	2
2012	14	3
2008	18	4
2004	22	5
2000	26	6
1996	30	7
1992	34	8
1988	38	9

The pattern (real age is always four times the actual-birthday count, plus two) falls straight out of the four-year leap cycle, not out of these particular birth years. 2026 sits two years past the nearest leap year, 2024, so every leap-day baby’s ordinary age is two years ahead of a clean multiple of four, no matter which leap year they were born in. Come the next leap year, 2028, that offset drops back to zero and everyone in the table gains one more actual Feb 29 birthday at once.

10—Units

Age in years, months and days is not age in days

There are two ways to express how old someone is, and they are not interchangeable representations of the same number.

Total elapsed days is unambiguous. Count the days between two dates and you get an integer that everyone agrees on, leap days included. It is what medical dosing, incubation periods and astronomical calculations use, precisely because it carries no calendar assumptions at all.

Years-months-days is what people actually say, and it is only well-defined once you fix the borrowing rules discussed earlier. “3 years, 2 months and 14 days” describes a walk across a calendar, not a quantity: the same phrase covers a different number of real days depending on which months it crossed.

The reason age calculators usually show both is that neither answers every question. The years-months-days figure is the one people came for; the day count is the one that survives being compared to another day count.

11—Wrapping up

The takeaway

Exact age math is simple in principle (years, months, and days since a starting date) but correct only if the months-and-days step uses real calendar month lengths, including leap years, instead of a fixed approximation. That’s the entire difference between an age calculator that quietly gives you a slightly wrong answer near month boundaries, and one that matches what you’d get counting it out by hand.

This kind of calendar-accurate math also matters beyond curiosity about your own age. Anything that depends on a precise duration between two dates (eligibility windows tied to a minimum age, warranty periods measured from a purchase date, or interest that accrues based on days elapsed) inherits the same borrow-a-month and leap-year edge cases described above. A tool that quietly rounds those cases to “close enough” can produce an answer that’s off by a day in exactly the situations where a day matters most, like a cutoff date for enrollment or eligibility. Building the calculation the calendar-accurate way from the start avoids that entire category of quiet, hard-to-notice error.