ASCII vs Binary STL: What's the Difference?

Every STL file describes a model the same way: a list of triangles. But the format allows two encodings of that list, and the choice makes a dramatic practical difference - the same model can be 5 MB or 25 MB depending on which variant saved it.

ASCII STL: human-readable, huge

An ASCII STL is plain text. It literally spells out every triangle: 'facet normal', 'outer loop', three 'vertex' lines with decimal coordinates, 'endloop', 'endfacet' - seven lines of text per triangle. You can open it in any text editor and read it, which is genuinely useful for debugging exporters or teaching the format. The price is bulk: all those characters typically make ASCII files four to five times larger than binary, and slower to parse.

Binary STL: compact, standard

A binary STL packs each triangle into exactly 50 bytes: a normal vector, three vertices as 32-bit floats, and a two-byte attribute field. An 80-byte header and a triangle count lead the file, so its size is perfectly predictable: 84 bytes plus 50 per triangle. This is what virtually all modern software writes by default, and what our converters output.

Does the geometry differ? No - with one footnote

Both variants store the same triangles, so converting between them is lossless in structure. The footnote: coordinates in both are 32-bit floats, so there is no precision difference either - despite ASCII showing you more decimal places, they encode the same values.

When to use which

• Binary: printing, sharing, archiving - effectively always
• ASCII: inspecting a file by eye, teaching, or debugging an exporter you are writing
• Received a suspiciously large STL? It is probably ASCII - re-saving it as binary is the single easiest fix in our file size guide