Count the sig figs in a number, or round to exactly N of them - handled correctly even on the edge cases (like 9.995) that trip up calculators doing the rounding in ordinary floating-point math.
| Rule | Example | Result |
|---|---|---|
| All non-zero digits count | 347 | 3 sig figs |
| Zeros between digits count (captive) | 1002 | 4 sig figs |
| Leading zeros never count | 0.0062 | 2 sig figs |
| Trailing zeros count only with a decimal point | 2.500 vs 2500 | 4 sig figs vs 2 |
The most common mistake is with whole numbers like 2500: without a decimal point, those trailing zeros are just placeholders holding the number's size, not measured digits - so 2500 has 2 sig figs, but 2500. (with an explicit decimal point) has 4, and 2.500×10³ also has 4. This ambiguity is exactly why scientists prefer scientific notation for numbers with trailing zeros.
Why this handles 9.995 correctly: rounded to 3 sig figs, the right answer is 10.0 - but 9.995 can't be stored exactly in binary floating point, and is actually held internally as 9.994999999999999. This tool rounds on the digits you actually typed instead of the imprecise stored float, so it gets tricky cases like this one right.