Matrix Calculator
Add, subtract, multiply, invert and transpose matrices up to 4×4 — plus RREF with rank, trace, and 2×2/3×3 eigenvalues, with every inverse automatically verified against the identity matrix.
⚡ Quick Tips for Working With Matrices Correctly
Features
Add, subtract, multiply, scalar multiply, transpose, determinant, inverse, trace, RREF and eigenvalues.
Every core operation generalizes to 4×4, not just the 2×2/3×3 limit most free calculators impose.
Every inverse is automatically multiplied back against A to confirm it reproduces the identity matrix.
Full Gauss-Jordan row reduction, with the resulting rank shown as the headline result.
Exact 2×2 and 3×3 eigenvalues, including complex-conjugate pairs when they occur.
One-click presets for every matrix, so testing a formula never means typing nine values by hand.
Why Ten Operations Instead of the Usual Six
Most free matrix calculators cover addition, subtraction, multiplication, determinant, inverse and transpose, and stop there — which is enough for a homework problem but not for a system of equations that needs to know whether it even has a unique solution, or a physics problem that needs eigenvalues to analyze stability. Scalar multiplication, trace, RREF with rank, and eigenvalues aren't exotic additions; they're the next four operations anyone doing real linear algebra reaches for once the basics are covered, which is why this tool treats all ten as equally first-class rather than bolting a few on as an afterthought.
Take a 3×3 coefficient matrix from a three-variable system of equations. The determinant tells you at a glance whether a unique solution exists at all. If it's zero, RREF mode row-reduces the same matrix and reports its rank directly — instantly showing which rows are redundant, rather than leaving that as a mystery. If the determinant is non-zero, computing the inverse and multiplying it against the constants vector solves the system outright, with the built-in verification confirming the inverse is genuinely correct before you trust it.
What Eigenvalues Actually Tell You
An eigenvalue answers a specific question: for this matrix, is there some direction that only gets stretched or shrunk — never rotated — when the matrix is applied to it? A 2×2 matrix always has two such values (possibly a complex-conjugate pair, possibly repeated); a 3×3 matrix has three. Real, distinct eigenvalues describe a transformation that stretches space along independent axes with no rotation involved; complex eigenvalues signal that some rotation is baked into the transformation itself — a rotation matrix in 2D, for instance, always produces a complex-conjugate pair, since rotation has no real direction that stays fixed.
This matters well beyond a linear algebra course. In mechanical and structural engineering, the eigenvalues of a stiffness matrix reveal a system's natural vibration frequencies — critical for avoiding resonance failures in bridges and aircraft wings. In population and epidemic models, the largest eigenvalue of a growth matrix determines whether a population trends toward growth, decline, or a stable equilibrium over time. In data science, Principal Component Analysis ranks a dataset's dimensions by the eigenvalues of its covariance matrix, keeping only the directions where the data actually varies the most.
The Trust Problem Every Matrix Calculator Has
A wrong inverse looks exactly like a right one — nine numbers in a grid, no obvious red flag. This calculator closes that gap by immediately multiplying the computed inverse back against the original matrix and comparing the product to the identity matrix, cell by cell, reporting the largest deviation found. A confirmed result near zero means the inverse has been mathematically proven correct for that specific input, not just computed by a formula that's assumed to be right. For a singular matrix, the same check is what makes the "no inverse exists" message trustworthy — it isn't guessing based on a near-zero determinant threshold alone.
Who Actually Reaches for a Matrix Calculator
Linear algebra students checking homework across every standard operation, not just the multiplication problems most textbook chapters lead with. Engineering and physics students solving small systems of equations from circuit analysis or structural load calculations, where the coefficient matrix rarely stays a clean 2×2. Computer graphics and game developers verifying transformation matrices by hand — rotation, scaling and translation are each single matrices multiplied together, and a quick inverse check confirms a transform can be undone correctly. Data science students working through the normal equations for linear regression, (XᵀX)⁻¹Xᵀy, which is a transpose, two multiplications and an inverse in sequence. For the individual arithmetic behind any of these formulas, the Scientific Calculator handles the scalar steps directly.
Two Calculations Worth Walking Through
Solving a system by inverse: the system 2x + y = 5, x + 3y = 10 becomes the matrix equation Ax = b with A = [[2,1],[1,3]]. det(A) = 2×3 − 1×1 = 5, so an inverse exists: A⁻¹ = (1/5)×[[3,−1],[−1,2]] — each entry a fraction over that same determinant, which the Fraction Calculator can simplify further if the numbers don't divide evenly. Multiplying A⁻¹ by b = [5,10] gives x = 1, y = 3 — solved directly through the inverse rather than substitution or elimination by hand.
Checking rank before trusting a system: three equations x+y+z=6, 2x+2y+2z=12, x−y=0 look like three independent constraints, but the second equation is just the first one doubled. RREF mode reduces this coefficient matrix and reports rank 2, not 3 — confirming only two of the three equations are actually independent, which means the system has infinitely many solutions rather than one unique answer.
Why Browser-Based Beats Manual Row Reduction
Gauss-Jordan elimination by hand on a 4×4 matrix is tedious and error-prone — a single arithmetic slip in an early row propagates through every step after it. This tool runs every operation, including the full RREF row reduction and the inverse verification pass, entirely client-side in JavaScript, with nothing entered ever transmitted to a server — safe for coursework, research data or proprietary engineering matrices kept private. For equations that reduce to a single quadratic rather than a full matrix, the Quadratic Solver is the more direct tool.
Where This Calculator Draws the Line
It handles square matrices up to 4×4 across all ten operations, with eigenvalues specifically limited to 2×2 and 3×3 for the exact closed-form reasons explained in the FAQ below. It doesn't perform rectangular (non-square) matrix operations, singular value decomposition, or eigenvalues for 4×4 and larger — those need iterative numerical methods better suited to dedicated software like NumPy or MATLAB. For statistics built on top of matrix math, like variance and standard deviation on a raw dataset, the Standard Deviation Calculator picks up where the underlying matrix arithmetic leaves off.