Every digit, not an approximation. 2 to the power of 200 has 61 of them, and you get all 61.
123,456,789,123,456,789 × 987,654,321,987,654,321 comes to exactly 121,932,631,356,500,531,347,203,169,112,635,269 — 36 digits. A normal calculator, limited to about 15–17 significant digits, would show roughly 1.219326×10³⁵: correct for the first several digits and silently wrong for the other twenty-plus.
Results are shown in full up to 2,000 characters and truncated beyond that — not because BigInt stops being exact, but because a browser tab rendering tens of thousands of digits stops being useful to read. 10,000! is the clearest case: it comes to exactly 35,660 digits, computed exactly by this page, but far past what the display shows in full. The exactness is real even when the display isn't; it is a rendering limit, not a computation one.
A credit card number's check digit, computed via the Luhn algorithm, depends on every digit being exactly right — a number one digit off in the trillions place fails validation exactly as completely as one off by 1. RSA encryption leans on the same all-or-nothing exactness at a much larger scale: the numbers involved run past 600 digits, and there is no such thing as an encryption key that is "approximately" correct — the Factor Calculator looks at the flip side of that same scale problem, where multiplying two large numbers together is instant but reversing the product back apart is not. For arithmetic that doesn't need this — everyday decimals, percentages, square roots — ordinary floating point is the right tool; see the Basic Calculator or Scientific Calculator.
Most calculators store numbers as 64-bit floating point, which holds about 15–17 significant digits exactly. Beyond that, digits are silently discarded.
Ask a normal calculator for 2²⁰⁰ and you get 1.606938e+60 — correct to seven digits and wrong about the other fifty-four. The exact answer is 1606938044258990275541962092341162602522202993782792835301376, and this tool gives all of it.
It is not a rounding preference. The information is genuinely gone: the format has nowhere to keep it.
This uses BigInt, which is exact but integer-only. Division returns the whole-number quotient with the remainder available separately — 7 ÷ 2 gives 3 remainder 1, not 3.5. That is a deliberate trade: exactness at any size, in exchange for no fractions.
For decimals, ordinary floating point is fine — the Basic Calculator or Scientific Calculator will serve you better.
Because it stores values as 64-bit floating point, which holds roughly 15 to 17 significant digits. Anything beyond that cannot be represented, so it is dropped. The result is displayed in exponential form, correct at the front and wrong behind it.
No. It uses BigInt, which is exact but works only on whole numbers. Division gives the integer quotient, with the remainder available as a separate operation. For decimal arithmetic use the Basic or Scientific Calculator.
Far larger than anything you are likely to need — the limit is memory rather than a fixed digit count. Factorial is capped at 10,000 and exponents at 100,000 here purely to keep the page responsive.
To avoid exactly this loss of precision. Storing £10.99 as 1099 pence keeps every value exact, whereas decimal floating point cannot represent many everyday amounts precisely and the error accumulates across transactions.
Rounding is the deliberate, visible version of losing precision on purpose — the Rounding Calculator covers the rules for doing that on ordinary decimal numbers, as opposed to the silent, unintentional precision loss this page exists to avoid entirely.
For very large or small magnitudes where an approximation is fine, the Scientific Notation Calculator is the right tool. Factorials also drive the Permutation and Combination Calculator, and prime work is in the Factor Calculator. Large values are often easier to read as hexadecimal, where every digit is exactly four bits — the hex converter does the conversion and the arithmetic.