π is exactly 22/7.
22/7 = 3.142857…, about 0.04% too big. It’s a handy approximation, not the true value.
Wrap a string around any circle, then compare it with the distance straight across the middle. The string is always a little more than three times as long. That ratio is π (pi), and it is the same for every circle there is.
Its digits never end and never fall into a repeating pattern. Yet π turns up everywhere: in wheels and water tanks, planet orbits, sound waves, exam-mark statistics and the electricity in your home. This page lets you measure it, calculate it, practise with it and explore a million of its digits.
3.14159 26535 89793 23846 26433 83279 50288 41971 69399 37510…
Every circle has exactly the same shape, from a bangle to a satellite’s orbit. So the distance around a circle divided by the distance across it always gives the same number.
Three lengths describe a circle. The radius (r) runs from the centre to the edge. The diameter (d) runs straight across through the centre, so it is twice the radius. The circumference (C) is the distance all the way around.
Divide the circumference by the diameter and you always get 3.14159…, however big or small the circle. Mathematicians write this number as π (pi), the Greek letter for p, because it begins the Greek words for perimeter and periphery. Rearranging the definition gives the formula you will use most often.
π = C ÷ dthe definition
C = πd = 2πrcircumference
Drag the slider. The circle grows and shrinks, but three diameters always fit around the edge with the same small gap left over: 0.14159… of a diameter.
Find a few round things: a steel plate, a bangle, a bottle cap, a coin. Wrap a thread around each one, mark where it meets itself, then straighten it along a ruler. Next, measure straight across the widest part. Use the same unit for both. The first three rows are example measurements; replace them with yours.
| Object | Circumference | Diameter | C ÷ d | Remove row |
|---|
π doesn’t only measure the edge. It tells you how much space a circle covers, and it is the natural unit for measuring turns.
Cut a circle into equal slices, like a pizza or a roti, and lay them side by side, pointing alternately up and down. The slices from the top half carry half the edge along the top; the slices from the bottom half carry the other half along the bottom. So the new shape is about πr long (half of 2πr) and r tall.
The more slices you cut, the flatter the wavy edges get, and the closer the shape comes to a rectangle. A rectangle’s area is length × height, so the circle’s area is πr × r = πr².
With 12 slices the shape is already close to a rectangle πr long and r tall.
Degrees split a full turn into 360 parts, a number chosen by ancient astronomers. Mathematicians prefer a unit built into the circle itself. Take the radius and bend it around the edge: the angle it covers is 1 radian, about 57.3°.
How many radius lengths fit around half a circle? Exactly π of them. So 180° = π radians, and a full turn is 2π radians. That is why π appears in almost every formula about rotation, from gear wheels to the hands of a clock.
Drag the point around the circle or use the slider. The arc shows how many radius lengths the angle covers.
To convert, multiply degrees by π/180, or radians by 180/π.
No fraction is exactly equal to π, and its decimals run on forever without repeating. Here is what that means, and how close a fraction can get.
Some decimals stop, like 3/8 = 0.375. Others repeat a block of digits forever, like 1/7 = 0.142857 142857 142857… Every fraction behaves in one of these two ways. π does neither.
In 1761 Johann Heinrich Lambert proved that π is irrational: no fraction of two whole numbers equals it. In 1882 Ferdinand von Lindemann proved more. π is transcendental, meaning it is not the solution of any polynomial equation with whole-number coefficients, the way √2 solves x² = 2.
Lindemann’s proof settled a puzzle posed by Greek geometers over 2,000 years earlier, known as squaring the circle. With only a compass and an unmarked straightedge, it is impossible to construct a square with exactly the same area as a given circle.
What is the closest fraction to π if the bottom number can be no bigger than the limit you set? Slide to raise the limit.
355113
3.1415929203…
Highlighted digits match π: 3.14159265358979…
Write π as a continued fraction and it starts 3 + 1/(7 + 1/(15 + 1/(1 + 1/(292 + …)))), written [3; 7, 15, 1, 292, 1, 1, 1, 2, 1, 3, …] for short. Stopping early gives the best fractions in turn: 3, 22/7, 333/106 and 355/113. The unusually large 292 means the next improvement needs a denominator above 33,000. That is why 355/113, found by Zu Chongzhi in China around 480 CE, is still the most famous fraction for π after 22/7.
Mathematicians suspect that π is normal: each digit appears a tenth of the time, each pair of digits a hundredth of the time, and so on, so any finite string of digits turns up somewhere. The first million digits fit that pattern well, as the digit counter shows, but nobody has proved it.
People have been pinning down π since the first cities. Each age found a cleverer way to squeeze out more digits, and Indian mathematicians made some of the biggest leaps.
Showing all 29 events. Amber dots mark work from India.
c. 1900–1600 BCE
A clay tablet found at Susa compares a hexagon’s perimeter with the circle drawn around it, in a way that amounts to π = 3 + 1/8 = 3.125.
c. 1650 BCE
The scribe Ahmes finds a circle’s area by squaring 8/9 of its diameter. That works out to π ≈ 256/81 ≈ 3.1605.
c. 800–500 BCE
These rulebooks for building fire altars explain how to turn a square into a circle of the same area. Their constructions imply π ≈ 3.088.
c. 250 BCE
Working in Syracuse, Archimedes draws 96-sided polygons inside and outside a circle and proves that 223/71 < π < 22/7, or 3.1408 < π < 3.1429. It is the first method that can be pushed to any accuracy.
c. 150 CE
The astronomer Ptolemy of Alexandria uses 377/120 ≈ 3.14167 in his great astronomy book, the Almagest.
263 CE
In China, Liu Hui doubles the sides of polygons up to 3,072 and arrives at 3927/1250 = 3.1416.
c. 480 CE
Zu Chongzhi shows that 3.1415926 < π < 3.1415927 and gives the fraction 355/113, correct to six decimal places. No one does better for about 900 years.
499 CE
In the Āryabhaṭīya he writes, in verse: add 4 to 100, multiply by 8 and add 62,000, and you have the approximate circumference of a circle 20,000 across. That gives 62,832 ÷ 20,000 = 3.1416. He calls the value āsanna, meaning approaching, which suggests he knew it was not exact.
628 CE
In the Brāhmasphuṭasiddhānta, Brahmagupta uses √10 ≈ 3.162 as a practical value.
c. 1400
Mādhava of Sangamagrama, founder of the Kerala school, finds that π/4 = 1 − 1/3 + 1/5 − 1/7 + …, about 270 years before Gregory and Leibniz found it in Europe. With clever correction terms he computes π as 3.14159265359, correct to 11 decimal places.
1424
In Samarkand, Jamshīd al-Kāshī uses a polygon with 805,306,368 sides (3 × 2²⁸) to compute π to 16 decimal places, a record for about 170 years.
c. 1501
Nīlakaṇṭha Somayājī’s Tantrasaṅgraha records π = 3 + 4/(2·3·4) − 4/(4·5·6) + 4/(6·7·8) − …, a series that closes in on π far faster than Mādhava’s first one.
1593
François Viète writes the first exact formula for π in Europe: an endless product of nested square roots.
c. 1610
Ludolph van Ceulen spends much of his life on polygons with up to 2⁶² sides and reaches 35 decimal places. The digits were carved on his tombstone in Leiden, and Germans long called π the Ludolphine number.
1655
John Wallis discovers that π/2 = (2·2·4·4·6·6·…) ÷ (1·3·3·5·5·7·…), a product built from simple fractions.
1706
John Machin finds a fast arctangent formula and computes 100 decimal places. In the same year William Jones becomes the first to use the Greek letter π for the ratio.
1730s–1740s
Leonhard Euler adopts the symbol π and spreads it through his books. He also shows that 1 + 1/4 + 1/9 + 1/16 + … = π²/6, and his work gives the famous identity eiπ + 1 = 0.
1761
Johann Heinrich Lambert, working in Berlin, proves that π cannot be written as a fraction of whole numbers.
1873
After years of calculating by hand, William Shanks publishes 707 decimal places. In the 1940s D. F. Ferguson finds a mistake at place 528, so only the first 527 were right.
1882
Ferdinand von Lindemann proves π is transcendental. Squaring the circle with compass and straightedge is impossible.
1897
A bill based on an amateur’s faulty circle-squaring method passes Indiana’s House of Representatives unanimously. Purdue professor C. A. Waldo, visiting the statehouse, helps persuade the Senate to shelve it.
1914
Srinivasa Ramanujan publishes 17 series for 1/π. One of them adds about 8 correct digits per term. His ideas lie behind the formulas used for every modern record.
1949
One of the first electronic computers spends about 70 hours computing 2,037 decimal places.
1973
Jean Guilloud and Martine Bouyer pass 1,000,000 decimal places on a CDC 7600 computer.
1989
The brothers David and Gregory Chudnovsky, using their own formula that adds about 14 digits per term, pass 1 billion decimal places.
1995
David Bailey, Peter Borwein and Simon Plouffe find a formula that produces any single hexadecimal digit of π without computing the digits before it.
2009
Fabrice Bellard computes about 2.7 trillion digits on a single desktop computer.
2019
Emma Haruka Iwao computes 31.4 trillion digits on Google Cloud, announced on Pi Day. In 2022 she reaches 100 trillion.
2025
Linus Media Group and KIOXIA reach 300 trillion digits in April. In November StorageReview computes 314 trillion on a single server that ran for about 110 days, the current record.
Since computers took over in 1949, the record has grown by a factor of about 150 billion. Each dot below is a new world record. The vertical scale grows a thousandfold at each gridline, so the steady climb means the record has grown exponentially.
Decimal places computed, logarithmic scale
| Date | Who | Decimal places |
|---|
Five ways to pin down π, from Archimedes’ polygons to the formula behind modern records. Every one runs live in your browser.
A polygon drawn inside a circle has a shorter perimeter than the circle; one drawn outside has a longer one. So π is trapped between them. Doubling the number of sides tightens the trap. With 96 sides, Archimedes proved π lies between 3.1408 and 3.1429.
Perimeters are divided by the diameter, so each figure is a lower or upper bound for π.
Starting from hexagons, each row doubles the sides using only square roots. No value of π is needed.
| Sides | Inside (lower bound) | Outside (upper bound) |
|---|---|---|
| 6 | 3.00000000 | 3.46410162 |
| 12 | 3.10582854 | 3.21539031 |
| 24 | 3.13262861 | 3.15965994 |
| 48 | 3.13935020 | 3.14608622 |
| 96 | 3.14103195 | 3.14271460 |
| 192 | 3.14145247 | 3.14187305 |
| 384 | 3.14155761 | 3.14166275 |
| 768 | 3.14158389 | 3.14161018 |
| 1,536 | 3.14159046 | 3.14159703 |
| 3,072 | 3.14159211 | 3.14159375 |
Archimedes rounded his square roots cautiously, so his published bounds, 223/71 and 22/7, are a little wider than the 96-sided row.
Scatter dots at random over a square that holds a quarter circle. The quarter circle covers π/4 of the square, so 4 × (dots inside ÷ all dots) estimates π. It works, but slowly: to gain one more correct digit you need about 100 times as many dots.
The estimate as dots are added
Inside the quarter circle Outside
The Comte de Buffon posed a strange question in 1733 and published the answer in 1777: if you drop needles on a floor made of parallel boards, how often does a needle cross a gap? When the needles are exactly as long as the boards are wide, the answer is 2/π of the time. So π ≈ 2 × needles dropped ÷ needles crossing.
In 1901 Mario Lazzarini claimed 355/113 from 3,408 drops. The result is so perfect that most mathematicians think he stopped at a lucky moment.
Crosses a line Doesn’t cross
An infinite series adds up endless terms that get smaller and smaller. Stop after a few terms and you have an approximation. Some series crawl towards π; Ramanujan’s and the Chudnovskys’ race there.
Highlighted digits match π.
Terms needed to get within 0.0000000001 of π, logarithmic scale
Your device can work out π exactly, digit by digit, using John Machin’s formula from 1706: π/4 = 4 × arctan(1/5) − arctan(1/239). It uses whole-number arithmetic on numbers thousands of digits long. Record-breaking runs use the Chudnovsky formula and a program called y-cruncher instead, but the idea is the same.
Pick a size to start. 50,000 digits takes a second or two.
π appears whenever something is round, turns, swings or travels in waves. It also turns up in places with no circle in sight.
Follow a point around a circle and plot its height. One full turn, 2π radians, draws one complete wave. That is why π sits inside every formula for sound, light, radio and alternating current.
A pendulum’s swing time depends only on its length: T = 2π√(L ÷ g), where g = 9.81 m/s². That is why pendulum clocks were the world’s best timekeepers for almost 300 years.
One swing there and back takes 2.01 s
Search for your birthday, count how often each digit appears, turn π into art, or listen to it.
Pick a date, or type any string of digits. The search runs through the first 1,000,000 decimal places, counting the first digit after the decimal point as place 1.
If π is normal, each digit from 0 to 9 should make up 10% of the digits in the long run. Count them and see how the shares even out as you include more digits.
Share of the first 1,000,000 decimal places, by digit
| Digit | Count | Share |
|---|
Each digit is a step in one of ten directions, 36° apart. The path wanders but never repeats.
Each digit plays a note from a pentatonic scale, 0 lowest and 9 highest. Because the digits never repeat, neither does the tune.
Ready at decimal place 1.
Check your working with the calculator, read the worked examples, then try fresh questions at three levels. Use π = 22/7 or 3.14, as in most school textbooks.
Enter any one measurement of a circle. The calculator finds the other three and shows every step.
Answers within 0.5% count as correct, so rounding to two decimal places is fine. You can type fractions such as 22/7.
A quick quiz on everything above, a memory challenge, and a word game where the lengths of your words spell out π.
Study the digits in groups of five, then type them from memory. One wrong digit ends the run.
Tip: say each group out loud like a phone number. Many people learn 14159 26535 89793 as three quick chunks.
Start typing. Your best so far: 0 digits.
In Pilish, the number of letters in each word matches the next digit of π: 3 letters, then 1, then 4, and so on. A 10-letter word stands for 0. Punctuation doesn’t count. Edit the example or write your own.
Records, oddities and celebrations from the world of π, and a few popular claims that don’t hold up.
Next Pi Day
14 March
Next Pi Approximation Day
22 July
15
Decimal places of π that NASA’s Jet Propulsion Laboratory uses for its most precise work, interplanetary navigation: 3.141592653589793.
37
Decimal places enough to work out the circumference of the observable universe, about 46 billion light-years in radius, to within the width of a hydrogen atom.
762
The decimal place where six 9s in a row begin. It is called the Feynman point, after the physicist Richard Feynman.
176,451
The decimal place where π’s opening digits, 314159, appear again.
604,341
India’s Independence Day written as 15081947 first appears at this decimal place. An 8-digit date has only about a 1% chance of appearing in the first million.
70,000
Decimal places recited from memory by Rajveer Meena at VIT University, Vellore, on 21 March 2015. He wore a blindfold for the nearly 10-hour recital, the Guinness World Record.
280
Decimal places recalled in one minute by 10-year-old Alberto Davila Aragon in 2025, another Guinness World Record.
314 trillion
Digits computed by StorageReview in November 2025 on a single server running for about 110 days. Printed 4,000 to a page in 400-page books, they would fill nearly 200 million volumes.
1706
The year William Jones first used the symbol π for this ratio. Euler made it popular in the decades after.
1897
Indiana’s House of Representatives passed a bill based on a wrong value of π. The state Senate set it aside and it never became law.
14 March
Pi Day, first celebrated in 1988 by physicist Larry Shaw at the Exploratorium in San Francisco. In 2019 UNESCO declared it the International Day of Mathematics. It is also Albert Einstein’s birthday.
τ = 2π
Some mathematicians argue that the circle constant should be τ (tau) ≈ 6.283, a full turn. Their Tau Day is 28 June.
Make your call on each claim, then read the explanation.
π is exactly 22/7.
22/7 = 3.142857…, about 0.04% too big. It’s a handy approximation, not the true value.
Every circle has the same ratio of circumference to diameter.
All circles are scaled copies of each other, so the ratio is always π. (This holds on a flat surface. On a curved surface like a globe, it changes.)
It has been proved that every possible string of digits appears somewhere in π.
It would follow if π is normal, which most mathematicians believe. But no one has proved it.
A US state once passed a law that changed the value of π.
Indiana’s House passed such a bill in 1897, but the Senate shelved it, so it never became law.
NASA needs only about 15 decimal places of π to navigate spacecraft.
JPL uses 3.141592653589793. The error from stopping there is far smaller than anything a spacecraft could notice.
π only matters when you are working with circles.
π appears in probability, statistics, waves and number patterns, such as 1 + 1/4 + 1/9 + … = π²/6.
People have known that π is a little more than 3 for nearly 4,000 years.
Babylonian and Egyptian scribes used values of 3.125 and about 3.16 around 1900–1600 BCE.
Computers have finally found the last digit of π.
There is no last digit. π is irrational, so its decimals never end. The record is 314 trillion and counting.
Everything you need for circle questions on one sheet, plus the key words from this page.
| Diameter | d = 2r |
|---|---|
| Circumference | C = 2πr = πd |
| Area | A = πr² |
| Radius from C | r = C ÷ 2π |
| Radius from A | r = √(A ÷ π) |
| Semicircle perimeter | πr + 2r |
|---|---|
| Semicircle area | ½πr² |
| Arc length | (θ ÷ 360) × 2πr |
| Sector area | (θ ÷ 360) × πr² |
| Ring area | π(R² − r²) |
| Cylinder volume | πr²h |
|---|---|
| Cylinder curved surface | 2πrh |
| Cylinder total surface | 2πr(r + h) |
| Cone volume | ⅓πr²h |
| Cone curved surface | πrl (l = slant height) |
| Sphere volume | ⁴⁄₃πr³ |
| Sphere surface | 4πr² |
| Hemisphere volume | ⅔πr³ |
| Half turn | 180° = π radians |
|---|---|
| 1 radian | ≈ 57.2958° |
| Everyday values | 22/7 ≈ 3.142857, 3.14 |
| Closer fraction | 355/113 ≈ 3.1415929 |
| π to 15 places | 3.141592653589793 |
The million digits on this page were computed with two independent programs that agree on every digit.