Wolfram Mathematica
Version 15.0.1 · Wolfram Research · Windows 10 / 11 (64-bit) · ranked #246 in the library
Symbolic and numerical computation in one environment, where an integral is solved exactly rather than approximated and the result drops straight into a plot in the same document.
Read the installation section below before you start. The order of the steps matters on this release.
About this release
Overview
Mathematica is a computation environment built on a symbolic language. That is the distinguishing property: expressions are manipulated as mathematical objects rather than immediately reduced to floating point numbers. An integral is solved in closed form where a closed form exists. An equation is rearranged symbolically. A series expansion is exact. Numerical methods are there when the symbolic route runs out, but they are the fallback rather than the starting point.
The notebook interface is the working surface. Input, output, plots, text, formatted mathematics and interactive controls all live in the same document, which serves as scratchpad, report and presentation at once. Attaching an interactive control to any expression takes one wrapper function, producing a slider or a selector that recomputes the result live, and that is how most exploratory work in it gets done.
Coverage is unusually wide for a single package. Linear algebra, calculus, differential equations, optimisation, statistics, probability, number theory, graph theory, geometry, signal and image processing, machine learning, and control systems are all in the same language with consistent conventions. Visualisation covers two and three dimensional plotting, vector fields, and interactive manipulation, all with the same styling system, and results export to the common document, image and data formats.
What it does
Feature breakdown
Exact symbolic computation
Expressions manipulated as mathematical objects, so integrals, sums, limits and equations are solved in closed form where one exists rather than approximated.
Notebook documents
Code, results, plots, formatted mathematics and prose in one document that works as scratchpad, report and slideshow without changing tools.
Interactive exploration
Wrap any expression to attach sliders, selectors or locators that recompute the result live, which is the fastest way to understand a parameterised model.
Very broad function library
Calculus, algebra, differential equations, statistics, optimisation, graphs, geometry, signal and image processing and machine learning in one consistent language.
Comprehensive visualisation
Two and three dimensional plotting, vector and field visualisation, statistical charts and geometry rendering, all sharing the same styling system.
Interface to other tools
Import and export across the common data, image and document formats, plus links to external languages and compiled code.
Changed in this build
Version 15.0.1
- Extended machine learning functions with more built in model types and better automatic selection.
- Improved symbolic solving for several classes of differential equations.
- Notebook interface performance improved on documents with many large plots.
- New geometry and computational region functionality.
- Faster numerical linear algebra on current multicore processors.
Installation
Follow these in order
- Disconnect from the network before installing.
- Extract the archive to a drive with at least 25 GB free.
- Pause real time protection.
- Run setup as administrator and install the offline data components when offered.
- Apply the included activation before launching.
- Open the application, evaluate a test expression, then restore protection.
Release notes
From whoever packed it
- Some functions normally reach a curated data service. Those specific functions will not work offline, everything computational does.
- Do not run the update check, it replaces components the activation modifies.
- Notebooks are plain text under the extension and can be kept in version control, which is worth doing on anything long lived.
System requirements
Minimum and recommended
| Operating system | Windows 10 or Windows 11, 64-bit |
|---|---|
| Processor | Multicore, computation scales across cores for many operations |
| Memory | 8 GB minimum, 16 GB or more for large symbolic work |
| Graphics | Any GPU with OpenGL support for three dimensional plotting |
| Storage | 20 GB with the offline data components installed |
Questions about this release
Answered before you ask
- How is it different from a numerical package?
- It works symbolically first. Numerical methods exist but exact solutions are the default where they are possible.
- Does it need a connection?
- Only for the curated data functions. All computation, plotting and notebook work runs offline.
- Can I use it without learning the language?
- Partly, through the natural language input, but the language is where the value is.
Download
Pick a mirror