# Tests & Benchmarks

- Link: https://opensourcecad.nl/en/tests-benchmarks/
- Published: 2026-10-22T18:44:27+02:00
- Author: Marijn van den Heuvel \+ AI

_This page is also available in Dutch: [Tests & benchmarks](https://opensourcecad.nl/tests-benchmarks/)._

I advise nothing I have not tried myself. This page holds the tests with which I
judge open-source CAD tools: reproducible, with a fixed method and honest conclusions.

It is deliberately a separate section: this is the _evidence_ behind my advice, 
not the story I come in with.

## What I test

 * **Modelling speed**: sketch, constraint, feature
 * **Assemblies**: mates, motion, bill of materials (BOM)
 * **CAM workflow**: toolpath generation and G-code quality
 * **Drawing automation**: TechDraw views, dimensioning and export
 * **STEP/IGES fidelity**: the round trip with commercial CAD
 * **Large assemblies**: 100+ parts and regeneration time

## How I measure

 * **Hardware:** AMD Ryzen 9 / 64 GB RAM / RTX 4080 (stated per test)
 * **OS:** Ubuntu 24.04 LTS (native) and Windows 11 where needed
 * **Metrics:** time to result, file size, number of errors, feature coverage
 * **Reproducible:** I publish test scripts and input files with the results

## Results

I publish results as soon as a test round is done. Is there a specific tool or situation
you want tested? I am happy to do that on request, and often a focused test in your
situation is worth more than a general benchmark.

[Ask for a test →](https://opensourcecad.nl/en/contact/)

## Reference

I keep the categories [BIM & Architecture](https://opensourcecad.nl/en/bim-architecture/)
and [Electronics & PCB](https://opensourcecad.nl/en/electronics-pcb/) as context,
but I do not test them actively, that is not my core field. The technology under
the tools is on [NURBS & modeling kernels](https://opensourcecad.nl/en/cad-kernels/).