Model 1 vs Model 2 QR codes

Two models, one survivor

"Model 1" and "Model 2" are the two generations of the ordinary square QR code. The terms surface in old documentation, library options and the occasional generator dropdown, and the practical answer never varies: everything is Model 2 now. What follows is why the distinction existed.

	Model 1 (1994)	Model 2
Largest version	14	40
Largest grid	73 × 73 modules	177 × 177 modules
Max numeric capacity	1,167 digits	7,089 digits
Alignment patterns	None	Yes, from version 2 up
Distortion tolerance	Poor on curves and tilt	Good
Status	Legacy, effectively extinct	The QR code, per ISO/IEC 18004

Model 1: the original

Model 1 is the code Denso released in 1994: three finder patterns, timing patterns, Reed-Solomon error correction, all the fundamentals already in place. Its ceiling was version 14: a 73 × 73 grid holding up to 1,167 numerals, about a sixth of what the format would eventually manage.

Its deeper limit was geometric. With only the corner finders and timing tracks to go on, a decoder had to assume the grid between them was regular. Print the code slightly stretched, photograph it at an angle, wrap it around a curve, and module positions drift from where interpolation expects them, errors that mount with symbol size.

Model 2: the fix that became the format

Model 2's defining addition is the alignment pattern: the small square-in-square marks distributed through the grid of every symbol from version 2 upward: one in a version 2 code, up to 46 of them by version 40. Each is a local surveying point, letting the decoder measure real distortion across the symbol and correct for tilt, curvature and lens perspective region by region rather than hoping the grid is straight.

That reliability gain is what made large symbols practical, and Model 2 raised the ceiling accordingly: 40 versions, a 177 × 177 grid, and the familiar 7,089-digit / 2,953-byte capacity maxima. The ISO/IEC 18004 standard is written around Model 2, with Model 1 retained only as historical material; JIS X 0510 in Japan tells the same story.

What this means in practice

• Choosing a "model" in a library or generator: it is Model 2, or the option is vestigial. UseQR, like every modern tool, produces Model 2 symbols exclusively.
• Reading old specs or academic papers that mention Model 1: historical context only. No current phone or scanner pipeline expects to meet one.
• Identifying a symbol in the wild: any code with a small square-in-square mark away from the three corners is Model 2 (version 2+). A version 1 Model 2 code has no alignment pattern either (at 21 × 21 it does not need one), so absence proves nothing, but at that size the distinction carries no practical consequence.

FAQ

What is the difference between QR code Model 1 and Model 2?

Model 1 is the 1994 original, limited to version 14 (73 x 73 modules, 1,167 digits) with no alignment patterns. Model 2 adds alignment patterns for distortion correction and extends to version 40 (177 x 177 modules, 7,089 digits).

Are Model 1 QR codes still used?

Effectively no. Model 2 superseded Model 1 decades ago, ISO/IEC 18004 is written around Model 2, and modern generators and scanners deal exclusively in it. Model 1 survives only in historical documentation.

Which model of QR code should I generate?

Model 2, which is what every current generator produces by default. If a tool offers a model choice at all, Model 2 is the correct option in every realistic scenario.

Why did Model 2 add alignment patterns?

To correct distortion. The small square-in-square marks act as surveying points across the symbol, letting decoders compensate for tilt, curvature and perspective, essential for the larger, denser versions Model 2 introduced.