Skip to content
UseQR

Reference tables

QR code data capacity table

The maximum a QR code can hold is 7,089 digits, 4,296 uppercase alphanumeric characters, 2,953 bytes or 1,817 kanji — all at version 40, error-correction level L. Raising to level H cuts each figure by roughly 57%.

Absolute maxima, version 40

Mode Level L Level M Level Q Level H
Numeric 7,089 5,596 3,993 3,057
Alphanumeric 4,296 3,391 2,420 1,852
Byte (UTF-8) 2,953 2,331 1,663 1,273
Kanji (Shift-JIS) 1,817 1,435 1,024 784

Bits per character by mode

Mode Bits Character set
Numeric 3.33 0-9
Alphanumeric 5.5 0-9, A-Z, space, $ % * + - . / :
Byte 8 anything (UTF-8 in practice)
Kanji 13 Shift-JIS double-byte

Alphanumeric mode does not include lowercase letters, ?, =, &, #, _ or ~. A single lowercase character or query separator forces the whole segment into byte mode.

The practical ceiling is much lower

A version-40 code is 177 modules across. For a phone camera at arm's length each module wants roughly 0.5 mm, giving a 9 cm square. Most applications should stay under about 300 bytes.

Typical payload sizes:

Payload Bytes Version at EC M Modules
Short URL 20 2 25
URL with a path 45 3 29
URL with full UTM tags 120 6 41
WiFi credentials 50 3 29
MeCard 95 5 37
vCard 350 12 65
UPI with note and reference 110 6 41
PIX BR Code 140 7 45

The cheapest capacity win

Switch modes. A URL written entirely in uppercase encodes in alphanumeric mode at 5.5 bits per character instead of 8 — roughly 45% more efficient. Schemes and hostnames are case-insensitive so HTTPS://EXAMPLE.COM is safe; paths usually are not, so only uppercase the scheme and host unless you know your server is case-insensitive.

FAQ

What is the maximum amount of data in a QR code?

7,089 digits, 4,296 uppercase alphanumeric characters, 2,953 bytes, or 1,817 kanji characters — all at version 40 with error-correction level L.

How many characters should a QR code actually hold?

Under about 300 bytes for anything scanned casually. The theoretical maximum produces a 177-module code needing roughly 9 cm of print to be readable.

Why does uppercase text fit better in a QR code?

Uppercase letters, digits and a few symbols use alphanumeric mode at 5.5 bits per character; anything else uses byte mode at 8 bits.

How much capacity does error correction cost?

Level H holds roughly 43% of what level L holds for the same version. Level M, the usual default, holds about 79%.

Try it — free, no signup