How many bits are in a byte?

Eight. A byte is 8 bits — eight 0-or-1 switches, giving 256 possible values. That part is settled. The more interesting question, and the one people actually type into Google, is why 8: a byte hasn't always meant eight, and the number that won was a deliberate choice made in the 1960s.

Why 8 bits — and not 6, 7 or 9?

The word byte was coined in July 1956 by Werner Buchholz, an engineer on IBM's Stretch (7030) supercomputer project. He spelled it with a y on purpose, so nobody would mistype it as "bit". At that point a byte's size was still up for grabs — Stretch itself could address bytes of different widths.

What locked it to eight was the IBM System/360, announced in 1964. It was the machine that made software portable across a whole product line, and it committed to the 8-bit byte. Eight is a power of two (2³), it divides cleanly into machine words, and 2⁸ = 256 values comfortably covered the characters of the day with headroom. Where IBM's System/360 went, the industry followed.

The 256-value sweet spot

The original ASCII standard (ASA X3.4-1963) is actually a 7-bit code — 128 characters, enough for the uppercase and lowercase alphabet, digits, punctuation and control codes. A byte's eighth bit was spare: it was used for parity checking, or to unlock a second 128 characters (accented letters, symbols) in "extended" sets like Latin-1. So a byte didn't just fit ASCII — it fit ASCII twice over. Turn all eight switches on — 11111111 — and you hit 255, the largest number one byte can hold; all off is 0, so its whole range is the 256 rungs from 0 to 255. Flip the switches in the builder above to "A" (65) and you're looking at the exact pattern a text file stores for that letter.

Bits vs bytes: the gotcha on your internet bill

A lowercase b means bits; an uppercase B means bytes. Internet plans are sold in bits per second (Mbps) because the numbers look bigger; files download in bytes per second (MB/s). Since there are 8 bits in a byte, a 100 Mbps connection tops out at 100 ÷ 8 = 12.5 MB/s. Nothing is broken — it's the 8 doing its quiet work.

Half a byte is a nibble

Split a byte in two and each 4-bit half is a nibble (sometimes spelled nybble — the byte/bite pun, continued). A nibble maps perfectly to one hexadecimal digit, 0–F, which is why one byte is always exactly two hex digits: 00 to FF. The builder's HEX readout is just its two nibbles read off in base 16.

Does a 64-bit computer have 64-bit bytes?

No. When a CPU, phone or operating system is called 32-bit or 64-bit, that number is the width of its registers and memory addresses — how big a chunk it handles or points to in one step — not the size of a byte. A byte is still 8 bits on a 32-bit smartwatch, a 64-bit laptop and a mainframe alike. The system width only sets how many bytes move at once: a 64-bit register holds 8 bytes lined up side by side.

When a byte wasn't 8 — and why "octet" exists

Before standardisation, byte width was a machine's own business. Various early computers used 6-bit, 7-bit or 9-bit bytes; the PDP-10's 36-bit word could be sliced into bytes of whatever size you declared. Because "byte" was historically ambiguous, networking specifications avoid it entirely and say octet — a word that can only ever mean exactly eight bits. It's also the answer to that stray crossword clue, "bits in a byte": EIGHT, or OCTET.

So the honest answer is two answers: today, always eight — but only because, sixty years ago, eight is the number that won.