Take the 2-minute tour ×
Programming Puzzles & Code Golf Stack Exchange is a question and answer site for programming puzzle enthusiasts and code golfers. It's 100% free, no registration required.

Your goal is to print (to the standard output) the largest number possible, using just ten characters of code.

  • You may use any features of your language, except built-in exponentiation functions.
    • Similarly, you may not use scientific notation to enter a number. (Thus, no 9e+99.)
  • The program must print the number without any input from the user. Similarly, no reading from other files, or from the Web, and so on.
  • Your program must calculate a single number and print it. You can not print a string, nor can you print the same digit thousands of times.
  • You may exclude from the 10-character limit any code necessary to print anything. For example, in Python 2 which uses the print x syntax, you can use up to 16 characters for your program.
  • The program must actually succeed in the output. If it takes longer than an hour to run on the fastest computer in the world, it's invalid.
  • The output may be in any format (so you can print 999, 5e+100, etc.)
  • Infinity is an abstract concept, not a number. So it's not a valid output.
share|improve this question
What do you exactly mean by "calculate". Also, If it takes longer than an hour to run on any computer in the world, it's invalid. is not objective. I could (theoretically) manufacture a computer that takes an hour to change one T-state –  user80551 Jun 13 '14 at 13:37
Does bit-shift count as an exponentiation operator since it is equivalent to * 2^x ? –  Claudiu Jun 13 '14 at 16:08
The fact that the time limit depends on the performance of the fastest computer in the world makes it impossible for us to determine the set of valid answers... I don't really like that –  David Z Jun 13 '14 at 19:45
@DavidZ: True. I'd like to see it changed to "1 hour on X computer", and also clarify the rules as to bit-shift, and whether other large built-in functions (e.g. ack in Wolfram) are allowed. I guess not the best question, but I enjoyed playing to the spirit of the rules –  Claudiu Jun 13 '14 at 20:47

41 Answers 41

up vote 5 down vote accepted

Wolfram Language


ack(9!,9!) = enter image description here

Output is in Arrow Notation.

share|improve this answer
This goes against the rule that says you can't get it from the Web. –  Kyle Kanos Jun 14 '14 at 2:13
@KyleKanos It doesn't my interpreter happens to be on the web. Like some other REPLs. –  Adam Speight Jun 15 '14 at 1:46
Not sure how that's different. Web access is web access, and it's not allowed as per the rules. –  Kyle Kanos Jun 15 '14 at 1:48
The Wolfram Language is available to use an Raspberry Pi and doesn't require web access. So isn't in violation of that rule. –  Adam Speight Jun 15 '14 at 2:15
I think it is very valid answer. Stupid reasoning for being on island and having no access to internet. 'No web' requirement is only to prevent printing output from a already available resource. All answers are valid as long as they 'run' a piece of code to calculate the output. Latest Visual Studio is coming that runs your code on the Azure cloud, so you want to restrict C#? Or you are ineligible if you have chromebook? –  microbian Jun 19 '14 at 17:08

Perl, >1.96835797883262e+18


Might not be the largest answer... today! But wait enough millennia and it will be!


To address some of the comments, by "enough millenia," I do in fact mean n100s of years.

To be fair, if the big freeze/heat death of the universe is how the universe will end (estimated to occur ~10100 years), the "final" value would be ~10214, which is certainly much less than some of the other responses (though, "random quantum fluctuations or quantum tunneling can produce another Big Bang in 101056 years"). If we take a more optimistic approach (e.g. a cyclic or multiverse model), then time will go on infinitely, and so some day in some universe, on some high-bit architecture, the answer would exceed some of the others.

On the other hand, as pointed out, time is indeed limited by the size of integer/long, so in reality something like ~0 would always produce a larger number than time (i.e. the max time supported by the architecture).

This wasn't the most serious answer, but I hope you guys enjoyed it!

share|improve this answer
Upvoted because I like the eventuality –  Tim Gostony Jun 13 '14 at 18:40
Won't time wrap around and return a small number at some point? Depends on whether it's 32-bit or 64-bit perl I guess –  Claudiu Jun 13 '14 at 20:48
1000 years ~ 3e10 seconds, so you'll still be hovering around 1e21 as your output. Perhaps if you waited 1e50 years, you might start competing against the other answers? –  Kyle Kanos Jun 14 '14 at 2:19
Do you pronounce this "time times time"? :-) (hi Timwi!) –  Pierre Arlaud Jun 16 '14 at 14:01
Is the time on death heat of the universe accounting for daylight savings? You could earn a few more seconds of existence. –  Alpha Jun 17 '14 at 17:35

Wolfram ≅ 2.003529930 × 1019728

Yes, it's a language! It drives the back-end of the popular Wolfram Alpha site. It's the only language I found where the Ackermann function is built-in and abbreviated to less than 6 characters.

In eight characters:

$ ack(4,2)


Or ≅ 2.003529930 × 1019728

ack(4,3), ack(5,2) etc. are much larger, but too large. ack(4,2) is probably the largest Ackermann number than can be completely calculated in under an hour.

Larger numbers are rendered in symbolic form, e.g.:

$ ack(4,3)

2↑²6 - 3 // using Knuth's up-arrow notation

The rules say any output format is allowed, so this might be valid. This is greater than 101019727, which is larger than any of the other entries here except for the repeated factorial.


$ ack(9,9)

2↑⁷12 - 3

is larger than the repeated factorial. The largest number I can get in ten characters is:

$ ack(99,99)

2↑⁹⁷102 - 3

This is insanely huge, the Universe isn't big enough to represent a significant portion of its digits, even if you took repeated logs of the number.

As large as it is, Graham's number surpasses anything expressible as an Ackermann number.

share|improve this answer
Not bad, but it's no match for my ack(4,2)+1 solution. –  user19057 Jun 13 '14 at 21:58
ack(4,2)+2, I WIN –  edvinas.me Jun 13 '14 at 22:07
I think you won. –  Anonymous Pi Jun 15 '14 at 17:34
@Kyle, Yimin There are lots of misleading statements in this post and the comments under it. This is not the Wolfram Language, and not what's used as the backend of W|A. It is simply "natural language" input to Wolfram|Alpha. In the Wolfram Language ack(4,2) is invalid input. W|A's natural language input shouldn't even count as a programming language. It is not Turing complete, and you can't write even basic programs in it such as Euclid's algorithm. For this reason I don't think this is a valid answer any more than a Google query would be. –  Szabolcs Jun 16 '14 at 13:48
The Wolfram Language is the programming language used in Mathematica, documented here. Wolfram|Alpha takes natural language as input, not the Wolfram Language. –  Szabolcs Jun 16 '14 at 13:49

CJam, 2 × 10268,435,457


This computes b, defined as follows:

  • a0 = 10

  • an = an - 12

  • b = 20 × a28

$ time cjam <(echo 'A28{_*}*K*') | wc -c
Real    2573.28
User    2638.07
Sys     9.46


This follows the same idea as Claudiu's answer, but it isn't based on it. I had a similar idea which I posted just a few minutes after he posted his, but I discarded it since it didn't come anywhere near the time limit.

However, aditsu's suggestion to upgrade to Java 8 and my idea of using powers of 10 allowed CJam to calculate numbers beyond the reach of GolfScript, which seems to be due to some bugs/limitations of Ruby's Bignum.

How it works

A    " Push 10.                                                          ";
28{  " Do the following 28 times:                                        ";
  _* " Duplicate the integer on the stack and multiply it with its copy. ";
}*   "                                                                   ";
K*   " Multiply the result by 20.                                        ";

CJam, ≈ 8.1 × 101,826,751


Takes less than five minutes on my machine, so there's still room for improvement.

This computes a20, defined as follows:

  • a0 = 20

  • an = (n × an - 1)2

How it works

KK,   " Push 20 [ 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 ]. ";
{     " For each integer in the array:                                 ";
  )*  " Increment it and compute the its product with the accumulator. ";
  _*  " Multiply the result with itself.                               ";
share|improve this answer
Haha, is that Kirby? :) –  aditsu Jun 13 '14 at 15:28
Could you describe what this is doing? –  Kyle Kanos Jun 13 '14 at 17:59
Thanks! That's pretty cool. –  Kyle Kanos Jun 13 '14 at 18:34
FYI, I ran some tests to check BigInteger performance - I calculated 100000! and converted to string. Results: java 6: 21 sec calculation, 44 sec toString; java 7: 14 sec calculation, 42 sec toString; java 8: 5 sec calculation, 1 sec toString! –  aditsu Jun 13 '14 at 20:27
@aditsu: Upgrading Java reduced the runtime from 5 minutes to 2 seconds! –  Dennis Jun 13 '14 at 21:53

Python2 shell, 3,010,301 digits


Calculating the length: Python will append a "L" to these long numbers, so it reports 1 character more than the result has digits.

>>> len(repr( 9<<9999999 ))

First and last 20 digits:

40724177878623601356... ...96980669011241992192
share|improve this answer
Darn it! This showed up whilst I was writing the same answer –  James_pic Jun 13 '14 at 15:05
That's really just 9 * 2**9999999, so one could argue it uses exponentiation. –  Dennis Jun 13 '14 at 15:29
Wait, ***built-in*** exponentiation functions are not allowed so this might slip under the rules. +1 –  user80551 Jun 13 '14 at 15:53
Mathematical way of computing the length: floor(log10(9 * 2**9999999))+1 –  Quincunx Jun 13 '14 at 18:09
9<<(9<<99) is a lot bigger, if it terminates. –  Keith Randall Jun 14 '14 at 23:16

Python 3, 9*2^(7*2^33) > 10^18,100,795,813

9*2^(2^35) > 10^10,343,311,894

Edit: My new answer is:


Old answer, for posterity:


Ten characters exactly.

I am printing the number in hex, and

You may exclude from the 10-character limit any code necessary to print anything. For example, in Python 2 which uses the print x syntax, you can use up to 16 characters for your program.

Therefore, my actual code is:


Proof that it runs in the specified time and generates a number of the specified size:

time python bignum.py > bignumoutput.py

real    10m6.606s
user    1m19.183s
sys    0m59.171s
wc -c bignumoutput.py 
15032385541 bignumoutput.py

My number > 10^(15032385538*log(16)) > 10^18100795813

3 less hex digits than the above wc printout because of the initial 0x9.

Python 3 is necessary because in python 2, 7<<33 would be a long, and << doesn't take longs as inputs.

I can't use 9<<(1<<36) instead because:

Traceback (most recent call last):
  File "bignum.py", line 1, in <module>

Thus, this is the largest possible number of the form a<<(b<<cd) printable on my computer.

In all likelihood, the fastest machine in the world has more memory than I do, so my alternate answer is:


9*2^(9*2^99) > 10^(1.7172038461*10^30)

However, my current answer is the largest anyone has submitted, so it's probably good enough. Also, this is all assuming bit-shifting is allowable. It appears to be, from the other answers using it.

share|improve this answer
So far this looks to be the largest answer by far. It's a 10,000,000,000 digit number, after all. –  nneonneo Jun 14 '14 at 15:09
@nneonneo : I believe my solution is larger :) –  Zaid Jun 15 '14 at 11:37

Any language with short enough constant names, 18 digits approx.


I would post this as a PHP answer but sadly M_PI makes this just a little too long! But PHP yields 8.0839634798317E+17 for this. Basically, it abuses the lack of absolute precision in PI :p

share|improve this answer
Can't you do tan(PI/2)? –  user80551 Jun 13 '14 at 15:56
@user80551 Could do, but I just noticed that I could make use of that last remaining character in the limit to boost my result by 99 times. –  Niet the Dark Absol Jun 13 '14 at 15:58
Too long in Common Lisp: (/ 99(sin pi)) (14 characters). Error in TI-83: 99/sin(π (8 characters) causes division by zero. Works in LibreOffice Calc: =99/SIN(PI( (10 characters, not counting =) computes 808423047055000000. LibreOffice auto-inserts the last two )) in =99/SIN(PI()). –  kernigh Jun 13 '14 at 20:12
It's depending on the precision of the language. In PHP 14 decimal digits is a common value. –  kenorb Jun 14 '14 at 9:42
@kernigh It doesn't matter in this case, but would 99/sin(π be five bytes or eight? I know TI-BASIC stores a bunch of commands as single bytes; is sin( one of them? –  WChargin Jun 15 '14 at 0:33

K/Kona: 8.977649e261 1.774896e308

  • !170 creates a vector of numbers from 0 to 169
  • 1.6+ adds one to each element of the vector & converts to reals (range is 1.6 to 170.6)
  • */ multiplies each element of the array together

If Kona supported quad precision, I could do */9.+!999 and get around 1e2584. Sadly, it doesn't and I'm capped to double precision.

old method

  • !99 creates a vector of numbers from 0 to 98
  • 9+ adds 9 to each element of the vector (now ranges 9 to 107)
  • 9.* multiplies each element by 9.0 (implicitly converting to reals, so 81.0 through 963.0)
  • */ multiplies each element of the vector together
share|improve this answer

Powershell - 1.25371912013172E+35


Multiplies 99 Petabytes times 999 Petabytes.

share|improve this answer
98901 sq petabytes? Is that some sort of method to measure the surface bit density of (future high-capacity)hard-disks? –  user80551 Jun 14 '14 at 13:33
+1 for PowerShell! –  Mike Shepard Jun 17 '14 at 3:20
Oh, nice. Didn't know pebibytes were possible by now. I always thought it stopped at tebibytes. –  Joey Jun 17 '14 at 17:40
@Joey Now if MS would only hurry up and add yobibytes, the answer could be even better. –  Rynant Jun 17 '14 at 18:26
I could add that to Pash if it helps ... ;-) –  Joey Jun 17 '14 at 21:45

J (((((((((9)!)!)!)!)!)!)!)!)

Yeah, that's a lot. 10^(10^(10^(10^(10^(10^(10^(10^6.269498812196425))))))) to be not very exact.

share|improve this answer
Does it actually output the number? Is there a free J compiler out there? –  Kyle Kanos Jun 13 '14 at 15:01
I did find a compiler, but after 20 minutes, !!9x hasn't printed anything to screen. I sincerely doubt that !!!!!!!!9x will ever be computed. –  Kyle Kanos Jun 13 '14 at 15:47
"If it takes longer than an hour to run on the fastest computer in the world, it's invalid." Not sure this would be valid since it doesn't run within an hour –  Claudiu Jun 13 '14 at 15:52
It took 70 minutes, but !!9x finally printed to screen. I'm impressed that it actually computed the value, but it still completely fails bullet 5. –  Kyle Kanos Jun 13 '14 at 17:04
@Quincunx: Likely true, however the condition for a valid answer is that the program must succeed in output; this answer completely fails that criterion. Sadly, the upvoters have neglected all the comments pointing this out and still upvote it over the CJam & Golfscript answers that are insanely huge and compute in reasonable time. –  Kyle Kanos Jun 13 '14 at 18:01


Without any tricks:

main = print -- Necessary to print anything
    $9999*9999 -- 999890001

Arguably without calculating anything:

main = print
    $floor$1/0 -- 179769313486231590772930519078902473361797697894230657273430081157732675805500963132708477322407536021120113879871393357658789768814416622492847430639474124377767893424865485276302219601246094119453082952085005768838150682342462881473913110540827237163350510684586298239947245938479716304835356329624224137216

Adapting Niet's answer:

main = print
    $99/sin pi -- 8.083963479831708e17
share|improve this answer
Third bullet says "Your program must calculate..." –  user80551 Jun 13 '14 at 19:51
floor(infinity) is a finite number? What the heck Haskell? –  nneonneo Jun 14 '14 at 15:23
1/0 != infinity, it's undefined. –  RubberDuck Jun 15 '14 at 17:39
Are you sure about that, @ckuhn203? In both GHC 7.6.3 and 7.8.2, I get isInfinite $ 1 / 0 -- True. As far as I can tell, IEEE 754 defines 1 / 0 as infinity. –  Taylor Fausak Jun 16 '14 at 14:30
Oh, you mean mathematically. I agree completely. But as far as programming with IEEE floating point (and Haskell in particular) is concerned, 1 / 0 == Infinity. –  Taylor Fausak Jun 16 '14 at 18:06

Golfscript, 1e+33,554,432


Computes 10 ^ (2 ^ 25), without using exponents, runs in 96 seconds:

$ time echo "10{.*}25*" | ruby golfscript.rb  > BIG10

real    1m36.733s
user    1m28.101s
sys     0m6.632s
$ wc -c BIG10
 33554434 BIG10
$ head -c 80 BIG10
$ tail -c 80 BIG10

It can compute up to 9 ^ (2 ^ 9999), if only given enough time, but incrementing the inner exponent by one makes it take ~triple the time, so the one hour limit will be reached pretty soon.


Using a previous version with the same idea:


Breaking it down:

8         # push 8 to the stack
{...}25*  # run the given block 25 times

The stack at the start of each block consists of one number, the current number. This starts off as 8. Then:

.         # duplicate the top of the stack, stack is now | 8 | 8 |
*         # multiply top two numbers, stack is now       | 64 |

So the stack, step by step, looks like this:

8 8
64 64
4096 4096
16777216 16777216

... etc. Written in math notation the progression is:

n=0, 8                     = 8^1  = 8^(2^0)
n=1, 8*8                   = 8^2  = 8^(2^1)
n=2, (8^2)*(8^2) = (8^2)^2 = 8^4  = 8^(2^2)
n=3,               (8^4)^2 = 8^8  = 8^(2^3)
n=4,               (8^8)^2 = 8^16 = 8^(2^4)
share|improve this answer
What's the deal with the two one's in front of the 2564 in your output? –  Kyle Kanos Jun 13 '14 at 16:19
@KyleKanos: That's not my output, it's the output of wc. I'll edit to make it clearer –  Claudiu Jun 13 '14 at 16:20
you could use wc -c to make the output clearer –  daniero Jun 13 '14 at 16:29
Nice! I had the same idea a few minutes ago, but it runs much slower with CJam (which is surprising, since it's faster than GolfScript in general). –  Dennis Jun 13 '14 at 16:44
Minor improvement: 10{.*}25 delivers 33,554,434 digits and finishes in 90 second on my machine. I don't know why, but 10{.*}26* prints nothing. –  Dennis Jun 13 '14 at 19:31

Python - Varies, up to 13916486568675240 (so far)

Not entirely serious but I thought it would be kinda fun.

print id(len)*99

Out of all the things I tried, len was most consistently getting me large ids.

Yielded 13916486568675240 (17 digits) on my computer and 13842722750490216 (also 17 digits) on this site. I suppose it's possible for this to give you as low as 0, but it could also go higher.

share|improve this answer
Great idea but unfortunately it doesn't calculate anything. –  user80551 Jun 13 '14 at 14:54
I do believe the *99 part invokes a calculation. –  commando Jun 13 '14 at 14:55
Oh yes, /me feels stupid now. –  user80551 Jun 13 '14 at 14:59
If you use something shorter - say id(id) or id(0j), you can multiply by 999 –  gnibbler Jun 13 '14 at 15:10
Out of smaller named ones in docs.python.org/2/library/functions.html , vars consistently gives the highest value(but 4 chars) followed by sum. Use print(sorted([(id(x),x)for x in[id,len,max,min,str,int,ord,chr,sum,map,abs,all,any,bin,bool,eval,oct,vars,ite‌​r,list,set,repr,round,zip,type,pow,dict,dir,hex]])[::-1]) to check. –  user80551 Jun 13 '14 at 15:49

wxMaxima ~3x1049,948 (or 108,565,705,514 )


Output is

269146071053904674084357808139[49888 digits]000000000000000000000000000000

Not sure if it quite fits specs (particularly the output format one), but I can hit even larger:


Output is


That's roughly 108,565,705,514 which is significantly larger than most of the top answers and was computed in about 2 seconds. The bfloat function gives arbitrary precision.

share|improve this answer

Perl (> 1e+21,474,836,479) (on a 64-bit machine)

(inspired by darkajax's solution)


$ perl -e 'print ~1x(1<<30)' | wc -c


The code generates a huge number because the calculation directly increases the exponent, rather than the number itself.

  • ~1 is 18,446,744,073,709,551,614 (roughly 18 billion-billion) on 64-bit architectures
  • x is the repetition operator
  • 1<<30 is 1,073,741,824
  • ~1x(1<<30) is a number in which ~1 (18 bn bn) is repeated over a billion times

I've tested this on a Linux cluster and it takes a few minutes to run. One could go higher if needed, but I don't have the patience to wait for the result

If the repetition counter is too large for the memory available, Perl notifies with an Out of memory! error.

Given sufficient memory, one could theoretically do something like:


which is roughly


You don't need to run it to know that it should work, thanks to O::Deparse:

$ perl -MO=Deparse -e 'print ~1x~1x~1x9'
print 18446744073709551614 x 18446744073709551614 x 18446744073709551614 x 9;
-e syntax OK

Not bad for 10 bytes of code...

share|improve this answer
Your program must calculate a single number and print it. You can not print a string, nor can you print the same digit thousands of times. –  Dennis Jun 15 '14 at 14:58
@Dennis : ~1 is being evaluated, as is 1<<30. And it's not printing the same digit thousands of times, else the code would print something like 99999999..... –  Zaid Jun 15 '14 at 15:23
But it's not a single number. –  Dennis Jun 15 '14 at 15:35
@Dennis : Why do you say that? It is a single number that can be held in a scalar: perl -e '$x= ~1x(1<<30); print $x;' –  Zaid Jun 15 '14 at 16:23
You have to calculate a single number, but you just repeat it. That's basically string concatenation. –  Dennis Jun 15 '14 at 16:26

Mathematica, 2.174188391646043*10^20686623745


Ten characters exactly.

share|improve this answer
Is it technically calculating anything, and is it outputting all the digits and not just scientific notation? –  Yimin Rong Jun 13 '14 at 20:12
@Yimin: The output may be in any format (so you can print 999, 5e+100, etc.) –  edc65 Jun 13 '14 at 23:14
The value is not set at compile-time, but depends on the particular machine on which the command is run. I think it counts. –  Michael Stern Jun 13 '14 at 23:25

Haskell, 4950

Aww man, that's not a lot! 10 characters start after the dollar sign.

share|improve this answer
Why not just print? Also, 9/0.000001 is greater than sum[1..99]. –  Taylor Fausak Jun 13 '14 at 19:11
At that rate, we might consider 9 999 999 999 to be a lower bound on results. –  Cory Jun 13 '14 at 20:22
@TaylorFausak This answer is obviously not to be taken seriously! –  Flonk Jun 16 '14 at 10:23

Python shell, 649539 999890001

Beats Haskell, not really a serious answer.

share|improve this answer
9999999999 is larger, isn't it? –  MadTux Jun 13 '14 at 13:59
@MadTux There is an ambiguous restriction to calculate the answer. –  user80551 Jun 13 '14 at 13:59
If this answer is correct, then mine 9**9**9**9 is correct either, as it's arithmetic operator (not built-in func). –  kenorb Jun 14 '14 at 13:28
@kenorb It's built in exponentiation that's not allowed. –  user80551 Jun 14 '14 at 13:30

Befunge-93 (1,853,020,188,851,841)

Glad nobody has done Befunge yet (it's my niche), but dammit I can't find any clever trick to increase the number.


So it's 9^16.


Basically multiplies the value at the top of the stack with itself. So, value at the top of the stack goes:




Outputs the final value. I would be interested to see if anybody has any better ideas.

share|improve this answer

Python, ca. 1.26e1388




share|improve this answer

Matlab (1.7977e+308)

Matlab stores the value of the largest (double-precision) floating-point number in a variable called realmax. Invoking it in the command window (or at the command line) prints its value:

>> realmax

ans =

share|improve this answer

Wolfram Alpha (does a website count as a language)?

9! ! ! ! !



thanks to Cory for the tip that spaces work as well as parens.

share|improve this answer
I'm pretty sure this is the biggest number in the thread as of my post but Wolfram chokes on comparing it to other numbers, even ((99!)!)! > 4 never returns. –  gggg Jun 13 '14 at 22:27
This goes against the rule that says you can't get it from the Web. –  Kyle Kanos Jun 14 '14 at 2:13
Too many W|A's! –  Anonymous Pi Jun 15 '14 at 17:37
I disagree that this breaks the rules. It doesn't "read it from the web" which is against the rules, it does in fact "calculate a single number and print it" which is what is it supposed to to. The fact that the only available interface is a website doesn't mean it can't count as a programming language. –  gggg Jun 16 '14 at 15:23
It is a completely valid answer. –  microbian Jun 19 '14 at 17:53

JavaScript, more than 12 835 665 026 163 (but it reads a clock)

9*new Date

Reading from a clock is effectively taking an input. Feels like cheating. However, the technique seems to be well-received in the answers here.

If clocks are not allowed, then:

JavaScript, 186 025 771 008


Trigonometry takes too many bytes (Math.sin, Math.PI).

Bit-shift goes near the integer maximum, then multiplication yields a floating point.

share|improve this answer
Bit shift operation are limited to 32 bit integers. 60 == 28 used in a shift operation (try in console: [7<<28, 7<<60]) –  edc65 Jun 13 '14 at 23:07

Perl, 1.84467422290351e+26 (On a 64-bit machine)

perl -le "print ~0*9999999"
share|improve this answer
Your code was the inspiration for my answer. –  Zaid Jun 15 '14 at 14:25


$ txr -p '(mask 999)'
share|improve this answer
What is this doing? –  g.rocket Jun 16 '14 at 0:15
@g.rocket It makes a bit mask in which bit 999 is set. It takes multiple arguments, e.g (mask 1 2) yields 6. –  Kaz Jun 16 '14 at 2:12

C++ prints 18446744073709551615


ULLONG_MAX is 10 characters.

#include <iostream>
#include <limits.h>
using namespace std;
int main()
return 0;
share|improve this answer
And what value does that print? –  Kyle Kanos Jun 14 '14 at 2:15
@KyleKanos Fixed it, thanks –  bacchusbeale Jun 14 '14 at 5:07

Bash / shell

You may exclude from the 10-character limit any code necessary to print anything.

... you can use up to 16 characters for your program.

(6 chars for code + 7 chars for printing = 26 digits) = 6888888 digits

$ seq -s9 999999

This number is bigger, but people may argue that it's in the wrong format.

The output may be in any format (so you can print 999, 5e+100, etc.).

(8 chars for code + 8 chars for printing = 26 digits) = more than 1183888008 digits

$ seq -s9 99999999

It usually takes around a minute to print the number and it contains the scientific representation of it.

Some funny example:

(10 chars = 26 digits)

echo $$$$$$$$$$

See: What does $$ mean in the shell?

share|improve this answer
Your program must calculate a single number and print it. You can not print a string, nor can you print the same digit thousands of times. –  aditsu Jun 14 '14 at 14:55

Python 2.7 - 8794643931199480236 15616093818140822

print hash('zz')
share|improve this answer
Um, that's 12 chars. –  user80551 Jun 13 '14 at 19:48
Ehh valid point @user80551. Guess I was slightly distracted by the footy match on TV :-) –  willem Jun 14 '14 at 6:17

Scala, 263-1

Poor, poor Scala. Takes at least 8 characters to get a BigInt value, which doesn't leave enough room to actually make it big.

But with only 7 characters of (counted) code, we can print the largest possible positive Long:

share|improve this answer

Perl (10^1000000000 - 1)


Try it like that:

D:>perl -e "print 9x99999999" > bigNum.txt

D:>dir bigNum.txt
 Volume in drive C has no label.
 Volume Serial Number is F02E-11C0

 Directory of d:\

14.06.2014 г.  18:52        99 999 999 bigNum.txt
               1 File(s)     99 999 999 bytes
               0 Dir(s)  34 293 600 256 bytes free

Or if you're patient you can just execute perl -e "print 9x99999999" and enjoy almost 100MB of 9's printed to your console :)

share|improve this answer
Useful reference:wellho.net/mouth/254_x-operator-in-Perl.html –  Brendan Long Jun 17 '14 at 18:33

protected by Dennis Jun 14 '14 at 18:37

Thank you for your interest in this question. Because it has attracted low-quality answers, posting an answer now requires 10 reputation on this site.

Would you like to answer one of these unanswered questions instead?

Not the answer you're looking for? Browse other questions tagged or ask your own question.