To be fair, this is one of the least worse ones.
Their system is quite good actually. Not as good as a GitHub repo with all the tests made explicit. But still pretty good.
They don't have an actual online judge system, all problems simply have an integer or floating point solution and they just check that you've found the value.
The only metric that matters is who solved the problem first after publication, e.g.: projecteuler.net/fastest=454. The "language" in which problems were solved is just whatever the user put in their profile, they can't actually confirm that.
Project Euler problems typically involve finding or proving and then using a lemma that makes computation of the solution feasible without brute force. As such, they live in the intersection of mathematics and computer science.
List of just the solution values:
Code solutions by individuals:Basically no one ever had the patience to solve them all. What we need is a collaborative solution.
Problems are under CC BY-NC-SA: projecteuler.net/copyright
Once you solve a problem, you can then access its "private" forum thread: projecteuler.net/thread=950 and people will post a bunch of code solutions in there.
How problems are chosen:
projecteuler.net says it started as a subsection in mathschallenge.net, and in 2006 moved to its own domain. WhoisXMLAPI WHOIS history says it was registered by domainmonster.com but details are anonymous. TODO: sample problem on mathschallenge.net on Wayback Machine? Likely wouldn't reveal much anyways though as there is no attribution to problem authors on that site.
www.hackerrank.com/contests/projecteuler/challenges holds challenges with an actual judge and sometimes multiple test cases so just printing the final solution number is not enough.
The beauty of Project Euler is that it would serve both as a AI code generation benchmark and as an AI Math benchmark!
Bibliography:
Claude says he's from the UK and has a background in mathematics. Oxbridge feels likely. How I Failed, Failed, and Finally Succeeded at Learning How to Code says he started off on the ORIC computer, which is British-made, so he is likely British.
This was a registration CAPTCHA problem as of 2025:Python solution:
Among the first 510 thousand square numbers, what is the sum of all the odd squares?
s = 0
for i in range(1, 510001, 2):
s += i*i
print(s)Solution:
233168Solutions to the ProjectEuler+ version:
The original can be found with:
printf '1\n1000\n' | euler/1.pyA(x) = x + 1
Z(u)(v) = v
S(u)(v)(w) = v(u(v)(w))S
(S)
(S(S))
(S(Z))
(A)
(0)
S
(S)
(
S
(S(S))
(S(Z))
)
(A)
(0)
S
(S(S))
(S(Z))
(
S
(
S
(S(S))
(S(Z))
)
(A)
)
(0)
S
(Z)
(
S(S)
(S(Z))
(
S
(
S
(S(S))
(S(Z))
)
(A)
)
)
(0)
S(S)
(S(Z))
(
S
(
S
(S(S))
(S(Z))
)
(A)
)
(
Z
(
S(S)
(S(Z))
(
S
(
S
(S(S))
(S(Z))
)
(A)
)
)
(0)
)
S
(S)
(S(Z))
(
S
(
S
(S(S))
(S(Z))
)
(A)
)
(0)So we see that all of these rules resolve quite quickly and do not go into each other.
S however offers some problems, in that:C_0 = Z
C_i = S(C_{i-1})
D_i = C_i(S)(S)Calculate the nine first digits of:
D_a(D_b)(D_c)(C_d)(A)(e)Removing
D_a:S^i(Z)S)(S)(D_b)(D_c)(C_d)(A)(e)As mentioned at euler.stephan-brumme.com these tend to be harder, as they have their own judge system that actually runs programs, and therefore can test input multiple test cases against their reference implementation rather than just hard testing the result for a single input.
Goes only up to Project Euler problem 254 as of 2025, which had been published much much earlier, in 2009, so presumably they've stopped there.
Articles by others on the same topic
There are currently no matching articles.