Last updated:

How Much Data Can a QR Code Hold? - Inside Those Black-and-White Squares

About a 7-minute read

Paying at a convenience store, boarding a train, adding a friend on LINE. You use QR codes every day, but do you know how much data fits inside those tiny black-and-white squares? The answer: up to about 1,800 kanji characters. That is roughly 4.5 pages of Japanese manuscript paper packed into one small square.

Data Capacity of a QR Code

The figures below are the ceilings for version 40, the largest size, with error correction set to the weakest level (L). These are the standard values as of 2026. Real QR codes are built at much smaller versions, so they almost never come close to these numbers.

Data TypeMaximum CapacityEveryday Comparison
Numbers only7,089 digitsAbout 640 phone numbers
Alphanumeric4,296 charactersAbout 15 X (formerly Twitter) posts of 280 characters
Binary (8-bit)2,953 bytesAbout 3 KB of data
Kanji (Shift_JIS)1,817 charactersAbout 4.5 manuscript pages

Numbers-only mode fits 7,089 digits, but kanji drops to 1,817 characters. As explained in Characters vs. Bytes, kanji takes more bytes per character, so fewer characters fit in the same space.

QR Code Versions and Sizes

QR codes come in versions numbered 1 through 40, and a higher version stores more data. Each side starts at 21 modules for version 1 and grows by 4 modules per step, so the number of modules per side is 4 x version + 17. The capacities in the table below are numbers-only ceilings at the weakest error correction level (L).

VersionModule GridMax Digits (Numbers, Level L)Common Use
1 (smallest)21 x 2141Short URLs
537 x 37255Typical URLs
1057 x 57652Business card contact info
2097 x 972,061Long text
40 (largest)177 x 1777,089Bulk data (rarely used)

Version 40 is made up of 177 x 177 = 31,329 black-and-white modules. In practice, almost nobody uses version 40. The more data you pack in, the bigger and more complex the code becomes, making it harder for phone cameras to scan.

What Everyday QR Codes Actually Contain

Here is what the QR codes you meet every day are carrying, organized by use case. The sizes are rough estimates based on the length of the URLs and parameters seen in 2026, not values that the services publish as specifications (for some of these uses, no specification is public at all).

Use CaseData InsideApproximate Size
LINE friend requestLINE URL (about 40 characters)About 40 bytes
PayPay / QR paymentPayment URL + tokenAbout 100-200 bytes
Website URLURL stringAbout 30-100 bytes
Wi-Fi loginSSID + password + encryption typeAbout 50-150 bytes
Business card (vCard)Name, phone, email, addressAbout 200-500 bytes
Airline boarding passFlight number, seat, passenger infoAbout 100-300 bytes

Most QR codes you scan every day use less than 10% of the maximum capacity. A LINE friend request is about 40 bytes - just 1.4% of the 2,953-byte limit.

Error Correction in QR Codes

One of the coolest features of QR codes is that they still work even when part of the image is dirty or covered.

Error Correction LevelRecoverable ShareData That Fits at the Same SizeBest For
L (Low)About 7%The mostClean environments (screens)
M (Medium)About 15%Slightly lessGeneral printed materials
Q (Quartile)About 25%Considerably lessDirty environments
H (High)About 30%The leastQR codes with logos

These percentages are not the share of the area you can cover up. They describe how many of the codewords (the units the data is split into, 8 bits each) can be restored. At level H, the original data can still be recovered when roughly three codewords in ten are unreadable, which is what lets companies lay a logo over the center of a QR code. Note, though, that if the finder patterns (the three large squares in the corners) are damaged, the scanner cannot lock onto the code in the first place, so recovery never gets a chance. That keeps the area you can cover to the middle of the symbol. And because stronger error correction leaves less room for data at the same size, the choice is a trade-off between tolerance for dirt and how much you can store.

QR Codes vs. Barcodes

FeatureBarcode (1D)QR Code (2D)
Data capacityAbout 20 alphanumeric characters7,089 digits / 4,296 alphanumeric characters
Data directionHorizontal onlyBoth horizontal and vertical
Scan angleFront only360 degrees, any angle
Error correctionNoneYes (up to 30%)
Japanese supportNoYes

A barcode holds about 20 alphanumeric characters at most - just enough for a product ID such as a 13-digit JAN code. A QR code takes 4,296 alphanumeric characters, so on a like-for-like comparison it carries more than 200 times as much, and it can hold kanji and kana directly.

That gap is exactly why the QR code exists: barcodes were no longer big enough for the factory floor. It was announced in 1994 by Denso (the QR code business is handled by Denso Wave today), and the work was done by a two-person team led by Masahiro Hara. The name stands for Quick Response, because the design aimed at reading speed rather than sheer capacity. The square finder patterns in three corners are the heart of that: their black-and-white widths follow a 1 : 1 : 3 : 1 : 1 ratio, a pattern that hardly ever turns up elsewhere in print. A scanner can therefore pin down the position of the code no matter which direction it sweeps from, which is what makes reading from any angle possible.

QR Code Security Risks

RiskHow It WorksHow to Stay Safe
Fake QR codesA sticker placed over the real codeWatch for suspicious stickers
PhishingCode links to a fake websiteCheck the URL after scanning
MalwareCode links to a malicious app downloadDo not scan unknown QR codes carelessly

QR codes are convenient, but their weakness is that you cannot see what is inside before scanning. A written URL lets you spot a suspicious site, but a QR code hides it until you scan. Just like password length and security, balancing convenience and safety matters.

Share this article