News

MathWorks R2026b adds variant and embedded software testing tools

The MATLAB and Simulink release adds tools for managing model variants and testing embedded code, alongside updates for vision, point clouds and AMD SoC design.
MathWorks Polyspace Test Server interface showing code coverage information

MathWorks has released R2026b of MATLAB and Simulink, adding tools aimed at teams that build and test embedded software across multiple product variants.

Simulink Variant Manager gives engineers a way to organise and inspect model configurations as designs branch into different hardware or feature sets. Polyspace Test Server adds centralised testing and code coverage workflows for embedded software. The company has also added a Visual Inspection Toolbox and a Point Cloud Toolbox for image and 3D sensor data work.

Other changes include support in SoC Blockset for AMD devices, TOML files and JSON data dictionaries, plus PyTorch co-execution in MATLAB workflows. The breadth of the release means teams will need to check the individual product and licence requirements for features they intend to use.

For design groups maintaining several embedded variants, the practical question is whether variant management and test evidence can be handled more consistently across the model and generated or hand-written code. MathWorks lists the new capabilities in R2026b, but the announcement does not quantify time savings or test coverage gains for a particular project.

Continue reading