Loading File Size Converter…
931.3226 GiB
1,000,000,000,000 bytes exactly
Different base, same bytes: 1 TB is 931.3226 GiB, because one TB is 1,000,000,000,000 bytes while one GiB is 1,073,741,824. This is the whole of the missing-space complaint: the label on the box counts in thousands and the operating system divides by 1024.
Memory is addressed in powers of two, so RAM is sold in binary units and a 16 GB module really holds 16 GiB. Storage is sold in decimal units, which is why a 1 TB drive is a full terabyte of bytes and still shows as 931 GiB once a file manager divides by 1024 three times. Neither figure is wrong: one is written in powers of ten and the other in powers of two.
Network and bus speeds are quoted in BITS, not bytes: a 1 Gbit/s link moves 125 MB a second at best, and 8 Mbit/s of upstream is a megabyte a second. The lower-case b is doing all the work in that distinction, which is why a download never runs at the number on the router.
There are two systems of file-size units in daily use. The decimal one climbs in powers of a thousand — kB, MB, GB, TB, PB — and the binary one climbs in powers of 1,024, with an i wedged into each name: KiB, MiB, GiB, TiB, PiB. The IEC introduced those binary names in 1998 precisely because both ladders were being written with the same three letters, and a great deal of confusion has outlived the fix.
This converter shows the amount you type on both ladders at once, plus the bit units networks are sold in, so you can see which definition a number came from instead of guessing at it.
One kibibyte is 1,024 bytes against a kilobyte’s 1,000, so the first rung is only 2.4% apart. The gap compounds at every step: 4.86% at MB against MiB, 7.37% at GB against GiB, 9.95% at TB against TiB, and 12.59% by the time you reach petabytes.
That compounding is the answer to the oldest complaint in computing. A drive advertised as 1 TB holds 1,000,000,000,000 bytes, exactly as promised. Divide by 1,024 three times, as Windows does while labelling the result GB, and the same drive reads 931.32 — hence the missing 69 gigabytes that were never there. A 500 GB SSD reports 465.66, and a 2 TB drive reports 1.82 in the binary units the operating system prefers.
Nothing is being taken. The label and the file manager are each dividing correctly, one in powers of ten and the other in powers of two, and the whole disagreement is which power they chose.
Memory is addressed by binary digits, so its capacity comes in powers of two whether the label says so or not: a 16 GB module really does hold 16 GiB, or 17,179,869,184 bytes. Storage vendors have always sold in decimal, and the US class actions over that labelling ended in added disclosures rather than a changed definition, so 1 TB means a round trillion bytes.
Operating systems then split. Windows reports binary values under decimal names, which is where the mismatch becomes visible. macOS has reported decimal units since 2009, so the same drive appears larger there than on Windows with no difference in the hardware. Most Linux desktops report decimal for storage and binary in tools such as free and du, sometimes on the same screen.
Network and bus speeds are quoted in bits per second while files are measured in bytes, and eight bits to the byte means a factor most people apply in the wrong direction. A 1 Gbit/s link tops out at 125 MB/s. A connection advertised as 100 Mbit/s downloads at 12.5 MB/s at best, and around 11 MB/s once protocol overhead is deducted. An 8 Mbit/s upstream is one megabyte a second.
The only thing distinguishing the two is the case of a single letter: lower-case b for bits, upper-case B for bytes. That is why a download indicator never shows the number printed on the router, and why an internet plan sounds eight times faster than it copies files.
Both answers are in circulation, which is the problem. Strictly, a kilobyte is 1,000 bytes and 1,024 is a kibibyte, and that is how storage vendors and macOS count. A great deal of software still says KB when it means 1,024, so if a figure matters, check which base produced it rather than assuming.
Because 500,000,000,000 bytes divided by 1,024 three times is 465.66, and your operating system is showing binary units under a decimal name. Roughly 30 GB of the difference is arithmetic rather than lost space; a little more goes on the filesystem itself, which needs room for its own metadata.
No — memory is sold in binary units even when labelled GB, so a 16 GB module is 16 GiB, about 7.4% more bytes than 16 decimal gigabytes. It is the one case where the ambiguous label works in your favour, and it is why memory sizes are always powers of two.
About five and a half minutes at the theoretical 12.5 MB/s, and closer to six in practice. Divide the advertised megabits by eight for megabytes per second, then trim about 10% for protocol overhead before dividing the file size by what is left.
Use kB for 1,000 bytes and KiB for 1,024 when precision matters, in documentation, invoices or anything a customer might measure. KB with a capital K has no agreed meaning and is best avoided in writing where the difference could be argued about later.