Multiply π by √163 — the third link prints 40.109169991133 — and raise e to it. The screen prints 2.6253741264077⋅10+17, and the number behind that rounding is 262 537 412 640 768 743.99999999999925, which is an integer to twelve nines and no closer than that to any other whole number. The second link prints 640 320³ + 744, and the screen cannot tell the two apart, because the display has room for fourteen significant digits and the two numbers first disagree in the fifteenth.
The near miss is arithmetic, not a fluke of notation. For the nine values of d for which Q(√−d) has class number one, the modular function j takes an integer value, and e^(π√d) sits within a rounding of |j| + 744; for d = 43 that gives 884 736 743.99977, which the fourth link shows, and for d = 163 it gives the number above, with the shortfall coming from the next term of the q-expansion, about 7.5⋅10−13. The integer is 262 537 412 640 768 744 and the constant is below it. It has to be: e^(π√163) is (−1)^(−√163), and the Gelfond–Schneider theorem makes that transcendental, so no whole number is anywhere near enough to be it. The 640 320³ is not museum piece either — the Chudnovsky formula divides by it and earns fourteen correct digits of π per term.
Answer
2.6253741264077⋅10+17
How to read it
e^(π√163) on the workbench, shown to fourteen significant digits. The second link is 640 320³ + 744, the integer it is just short of; the third is the exponent, the fourth the weaker d = 43 case.
Source
Srinivasa Ramanujan's 1913 remark on e^(π√163); the explanation runs through the Heegner–Baker–Stark class-number-one theorem, and j((1 + √−163)/2) = −640 320³
Work it in the calculator
These links open the right calculator with this problem's numbers already typed in. The line above is what the screen should show — check it yourself.