QR code versions 1 to 40: sizes, capacity and how to choose

QR codes come in 40 versions. Version 1 is 21×21 modules and each version adds 4 modules per side, up to 177×177 at version 40, which holds up to 2,953 bytes. Generators auto-select the smallest version that fits the payload at the chosen error-correction level, and smaller scans better.

The 40 sizes

A QR code's version is its size class, nothing more. The rule is one line:

side length in modules = 4 × version + 17

Version 1 is 21 × 21; every step adds 4 modules per side; version 40 is 177 × 177: 31,329 modules. Capacity grows faster than area suggests, because the fixed overhead (finders, timing, format bits) is bought once: at version 1 function patterns consume 53% of the symbol, at version 40 under 9%.

Milestones (byte capacity at level M):

Version	Modules	Bytes at M
1	21 × 21	14
2	25 × 25	26
5	37 × 37	84
10	57 × 57	213
40	177 × 177	2,331

The ceiling at version 40, level L, is 7,089 numeric / 4,296 alphanumeric / 2,953 bytes. Full per-version figures are in the capacity table and the version and module table.

Structural features arrive on a schedule too: alignment patterns from version 2, the explicit version information field from version 7.

How the version is chosen

You normally do not choose it. Encoders (UseQR's included) take the payload and the error-correction level and select the smallest version whose capacity holds the encoded bit stream. Three inputs move the outcome:

Payload length, the dominant factor.
Encoding mode, 100 digits fit in far fewer bits than 100 arbitrary bytes.
EC level: raising L → H roughly halves data capacity at a given version, pushing the payload into a bigger symbol.

Measured with UseQR's encoder: a 71-character URL needs version 5 (37 × 37) at level M but version 8 (49 × 49) at level H. Same data, 44% more modules, each one smaller at the same printed size.

Why smaller is better

Every version step makes modules smaller at a fixed physical size, and module size is what limits scan distance and camera focus. A version 3 code scans from further away, survives worse printing and forgives more styling than a version 12 code at the same footprint. The practical ordering:

Shorten the payload (a short URL can drop several versions).
Set the EC level the environment needs, not more.
Let the version fall where it falls, then size the print for the distance.

Forcing a higher version than needed is occasionally useful (batches where every code must be the same size regardless of payload), and UseQR's API accepts a version parameter for exactly that. Forcing a lower one is impossible: if the data does not fit, generation fails rather than silently truncating.

FAQ

What do QR code versions mean?

Version is the symbol's size class, from 1 (21×21 modules) to 40 (177×177). Each version adds 4 modules per side and therefore more capacity. It has nothing to do with software versions or the age of the standard.

Which QR code version should I use?

The smallest one that fits your payload at your chosen error-correction level, which is what generators select automatically. Lower versions have larger modules at the same printed size, so they scan from further away and tolerate more damage.

How much data does a version 40 QR code hold?

At error-correction level L: 7,089 numeric digits, 4,296 alphanumeric characters, or 2,953 bytes. At level H that falls to 1,273 bytes. In practice codes that large are hard to scan from print and are rarely a good idea.

How do I find a QR code's version?

Count the modules along one edge and compute (count − 17) ÷ 4. A 29-module side is version 3. Decoders do the same measurement via the timing patterns, cross-checked against the version information field from version 7 up.