About this release
MATLAB is the environment that a great deal of engineering and scientific work is actually done in. The language is built around matrices, so the code on the screen reads like the mathematics on the page, and the workspace keeps every variable visible and inspectable while you work. For anyone developing an algorithm rather than shipping an application, that loop between writing a line and seeing what it produced is the whole point.
This release carries the toolbox set rather than the bare language. Signal processing, image processing, control systems, statistics and machine learning, optimisation, symbolic mathematics, curve fitting and the parallel computing toolbox all install from the same source, along with Simulink for block diagram modelling and simulation. That collection is what turns the environment from a calculator into a working platform for a discipline.
Visualisation is a first class part of the language rather than a library bolted on. Two and three dimensional plots, surfaces, volumes and animations come from single statements, every element of a figure is programmatically addressable, and the export path produces figures that go straight into a paper or a report at publication quality.
The apps and the code generation side cover what happens after the algorithm works. Interactive apps for classification, curve fitting, filter design and system identification let you work by hand and then generate the equivalent code, and the compiler and coder products turn a finished algorithm into a standalone program or into C code for an embedded target.
What it does
- Matrix language
- Syntax built around linear algebra so the code stays close to the mathematics.
- Full toolbox set
- Signal, image, control, statistics, optimisation, symbolic and parallel toolboxes included.
- Simulink
- Block diagram modelling and simulation for dynamic and embedded systems.
- Publication quality figures
- Plots, surfaces and animations from single statements with full programmatic control.
- Interactive apps
- Curve fitting, classification, filter design and system identification with code generation from the session.
- Live scripts
- Code, formatted text, equations and output in one runnable document.
Changed in this version
- Faster execution on high core count machines across the core numerical routines.
- New solvers added to the optimisation toolbox.
- Live editor gains better handling of very long output.
- Improved graphics performance on large three dimensional plots.
- Updated interfaces for calling code written in other languages.
System requirements
| Processor | Any 64 bit processor with AVX2 support, four cores or more recommended |
|---|---|
| Memory | 8 GB minimum, 32 GB for large models |
| Graphics | GPU with OpenGL 3.3 support, compute capable card for accelerated toolboxes |
| Disk | 40 GB free for a full toolbox install, solid state recommended |
| Display | 1920 by 1080 or higher |
Install order
- Unpack the archive or mount the image on a drive with plenty of free space.
- Run the installer and select only the toolboxes you actually need to keep the install manageable.
- Let the installer finish without interrupting it, since a partial toolbox install is hard to repair.
- Start the environment once and let the first launch build its path and cache.
Worth knowing before you start
A full toolbox install is very large, so check free space before starting.
Selecting fewer toolboxes shortens the install considerably and can be extended later.
File details
| Title | MathWorks MATLAB R2025b |
|---|---|
| Version | 25.2.0.2998904 |
| Publisher | MathWorks |
| Section | Engineering and Science |
| Edition | Full release with toolboxes |
| Size on disk | 24.7 GB |
| Platform | Windows 11, Windows 10 |
| Architecture | 64 bit |
| Languages | English, Japanese, Korean, Chinese, German |
| Mirrors carrying it | 7 |
| Added to the library | 10 months ago |
| Last refreshed | 1 week ago |