What Are KB, MB, and GB?
File sizes are byte counts wearing unit labels. A byte is the smallest addressable unit of storage — eight bits, enough for one ASCII character — and files range from a few bytes for a small text snippet to terabytes for a video library. The labels KB, MB, GB, and TB are shorthand for "a thousand bytes", "a million bytes", "a billion bytes", and "a trillion bytes" — or their 1024-based twins. The trouble, as anyone who has stared at a "931 GB" drive bought as 1 TB knows, is that the two counting systems do not agree, and both are everywhere. This converter exists to make that difference explicit instead of leaving you to guess which one your software is using.
Decimal vs Binary: The Whole Story
The decimal convention follows the SI prefixes: kilo- means 1000, mega- means a million, giga- a billion. It is the language of marketing and of most storage hardware — when a drive's box says 500 GB, it means 500,000,000,000 bytes. The binary convention comes from computing itself: memory and storage are naturally organized in powers of two, so early systems counted 1024 bytes as a "kilobyte", 1024² as a "megabyte", and so on. For decades software used the SI labels for binary amounts, which is why file managers worldwide report 1024-based sizes labeled KB/MB/GB. In 1998 the IEC introduced KiB/MiB/GiB/TiB to name the binary values, but adoption has been slow, and the ambiguity survives in every operating system today.
How the Converter Works
Pick a convention and the tool converts through bytes. Decimal mode multiplies by powers of 1000; binary mode by powers of 1024. From there, any source unit can meet any target unit: 5 GB decimal becomes 5,000 MB; 5 GiB binary becomes 5,120 MiB; and Auto mode converts whatever you enter to bytes and then picks the largest unit whose value stays at or above 1, so 1,500,000 bytes reads as 1.5 MB rather than 1,500 KB. Results are rounded to four decimal places for display while the internal math stays exact.
Step-by-Step: One File, Two Answers
Suppose a video file is exactly 5,000,000,000 bytes — the size of a "5 GB" camera card. In Decimal mode, entering 5 GB and converting to MB returns 5,000 MB, because 5 × 1000 = 5000. Now switch to Binary mode: the same 5,000,000,000 bytes divide by 1024 to 4,882,812.5 KB, then to about 4,768.37 MiB, or roughly 4.66 GiB. Same file, two correct answers, a 7.4% difference. That gap is exactly why the mode toggle matters: "how big is this file?" and "how much space does it take on my disk?" are the same question answered with different rulers.
A Short History of the 1024 Confusion
The conflict is as old as the byte itself. Early computers organized memory in powers of two because addresses are binary, and 1024 — being 2¹⁰, conveniently close to 1000 — got called a "kilobyte" almost immediately. Software followed hardware, and for decades the computing industry happily used SI labels for binary amounts. The trouble began when storage became consumer hardware: a manufacturer counting 1,000,000 bytes per megabyte and an operating system counting 1,048,576 both claim to sell and show "megabytes". In 1998 the IEC settled the naming — kibibyte, mebibyte, gibibyte for the 1024-based values, reserving kilobyte and friends for decimal — and while every serious technical reference now uses those names, the old habit of calling binary amounts MB and GB is still baked into most interfaces, which is why the ambiguity refuses to die.
What a Byte Actually Buys You
Having a feel for real-world sizes makes conversion results meaningful. A plain-text email or a short code file is a few KB. A typical web page with images is around 2–5 MB, a smartphone photo roughly 3–7 MB, and a minute of 1080p video about 100–200 MB. A full-length movie in streaming quality runs 2–5 GB, a modern video game 50–150 GB, and a professional video library crosses into terabytes. When you convert "4 GB to MB" and get 4,000 or 4,096 depending on mode, that ~2.4% difference decides whether a 4 GB memory card holds 1,000 or 1,024 of the same photos — small at the MB scale, and compounding as sizes grow.
Why Different Apps Show Different Sizes
Two file managers can honestly disagree about the same file. One rounds to the nearest whole unit and counts in binary; another keeps two decimals and counts in decimal. A 2,400,000-byte file is 2.4 MB by the decimal count and 2.29 MiB by the binary count — and if both labels print "MB", they look like different answers to the same question. This converter gives you the numbers both ways on purpose, and the Auto mode shows you the most readable unit so a size like 1,500,000 bytes presents as 1.5 MB instead of 1,500 KB.
Reference Table
| Unit | Exact bytes | Also equal to |
|---|---|---|
| 1 KB | 1,000 | ≈ 0.977 KiB |
| 1 MB | 1,000,000 | ≈ 0.954 MiB |
| 1 GB | 1,000,000,000 | ≈ 0.931 GiB |
| 1 TB | 1,000,000,000,000 | ≈ 0.909 TiB |
| 1 KiB | 1,024 | exact in both conventions |
| 1 MiB | 1,048,576 | exact in both conventions |
| 1 GiB | 1,073,741,824 | exact in both conventions |
| 1 TiB | 1,099,511,627,776 | exact in both conventions |
The KiB/MiB rows are identical in both conventions because binary mode is defined by those exact byte counts — the difference only appears when labels cross conventions. Whatever mode you pick, the byte count in the middle column is the number that never lies: convert everything to bytes when you need an apples-to-apples comparison between two systems that disagree.
Beyond Terabytes: Petabytes and Up
The prefix ladder does not stop at terabyte. Each step multiplies the previous unit by another thousand in the decimal convention and by 1024 in the binary one, giving the petabyte, exabyte, zettabyte, and yottabyte family:
| Unit | Decimal (×1000) | Binary (×1024) |
|---|---|---|
| PB / PiB | 10¹⁵ B · 1,000 TB | 2⁵⁰ = 1,125,899,906,842,624 B |
| EB / EiB | 10¹⁸ B · 1,000 PB | 2⁶⁰ = 1,152,921,504,606,846,976 B |
| ZB / ZiB | 10²¹ B · 1,000 EB | 2⁷⁰ = 1,180,591,620,717,411,303,424 B |
| YB / YiB | 10²⁴ B · 1,000 ZB | 2⁸⁰ = 1,208,925,819,614,629,174,706,176 B |
These units sound abstract, but they describe real infrastructure. A petabyte is roughly the storage of a serious media archive or scientific dataset — the sort of volume that lives in a data center rather than on a desk. Cloud providers quote their total capacity in exabytes, and industry estimates of the world's digital data — videos, backups, sensor readings, and everything else — run in the zettabyte range, with the numbers growing every year. The yottabyte, a septillion bytes, is still mostly a theoretical frontier: no single system manages that much data yet, but the prefix ladder is ready for it. The same decimal-versus-binary gap that makes a 1 TB drive show as 931 GiB applies at every rung, and it compounds — the two conventions drift apart by about 2.4% per step, so by the yottabyte scale they disagree by over 20%.
When Each Convention Applies
- Storage hardware (decimal): drive capacities, SSD sizes, memory cards, and most cloud storage quotas are quoted in 1000-based units.
- Operating systems (binary): Windows, macOS Finder, and Linux tools typically compute with 1024 while printing KB/MB/GB.
- RAM (binary): memory modules are 1024-based — a 16 GB stick is 16 GiB.
- Network and browsers (decimal): bandwidth in Mbps and most download meters use 1000-based units.
- APIs and specs (bytes): the unambiguous ground truth is always the raw byte count; convert everything to bytes before comparing.