What Is an Augmented Matrix?
An augmented matrix [A|b] combines the coefficient matrix A of a linear system with the constant vector b into a single matrix. The vertical bar (or a column separator) marks the boundary between coefficients and constants. For the system:
2x + y − z = 8
−3x − y + 2z = −11
−2x + y + 2z = −3
the augmented matrix is:
[ 2 1 −1 | 8 ]
[ −3 −1 2 | −11 ]
[ −2 1 2 | −3 ]
Row reducing this augmented matrix to RREF gives the solution directly — no back substitution required. Each pivot row corresponds to one solved variable. The RREF calculator above performs this automatically when you enter your augmented matrix. Enable "Augmented" mode to tell the calculator which column is the constant column.
How to Use This Augmented Matrix Calculator
- Choose your matrix dimensions. For n equations with n unknowns, select an n×(n+1) matrix. The last column will hold the constants.
- Enter the coefficients. For each equation, enter the coefficients in the first n columns. Enter the constant (right-hand side) in column n+1.
- Enable Augmented mode. Toggle the "Augmented" switch so the calculator treats the last column as the constants column b, displaying a separator bar.
- Click Calculate RREF. The calculator applies Gauss-Jordan elimination and shows every row operation.
- Read the solution. Each pivot row gives one variable. Free-variable columns correspond to parameters in infinite-solution cases. A row like [0 0 ... 0 | c] with c ≠ 0 means no solution.
The Three Solution Types
Every system of linear equations falls into exactly one of three cases, revealed immediately by the RREF of [A|b]:
Case 1: Unique Solution
The RREF of [A|b] has the form [I|x*] where I is the identity matrix and x* is the unique solution vector. Every variable column has a pivot. rank(A) = rank([A|b]) = n.
[ 1 0 0 | 2 ] → x = 2
[ 0 1 0 | 3 ] → y = 3
[ 0 0 1 | −1 ] → z = −1
Case 2: Infinitely Many Solutions (Free Variables)
Some variable columns have no pivot — these are free variables. Set each free variable to a parameter (t₁, t₂, …) and express the pivot variables in terms of them. rank(A) = rank([A|b]) < n. For a deeper treatment of free variables and the null space, see our null space guide.
[ 1 0 2 | 3 ] → x = 3 − 2t (z = t is free)
[ 0 1 1 | 1 ] → y = 1 − t
[ 0 0 0 | 0 ] → redundant equation (ignored)
Solution: (3−2t, 1−t, t) for any t ∈ ℝ
Case 3: No Solution (Inconsistent)
A row of the form [0 0 … 0 | c] with c ≠ 0 appears. This represents "0 = c", which is impossible. rank(A) < rank([A|b]).
[ 1 2 | 5 ]
[ 0 0 | 3 ] ← "0 = 3" — impossible. No solution.
These three cases correspond to the Rouché–Capelli theorem: unique solution when rank(A) = rank([A|b]) = n; infinite solutions when rank(A) = rank([A|b]) < n; no solution when rank(A) < rank([A|b]). See our solving linear systems guide for detailed coverage of all three cases.
Worked Example: 3×3 System with Unique Solution
Solve: 2x + y − z = 8, −3x − y + 2z = −11, −2x + y + 2z = −3.
Write as augmented matrix and apply Gauss-Jordan:
[ 2 1 −1 | 8 ]
[ −3 −1 2 | −11 ] → Gauss-Jordan elimination
[ −2 1 2 | −3 ]
RREF: [ 1 0 0 | 2 ] → x = 2
[ 0 1 0 | 3 ] → y = 3
[ 0 0 1 | −1 ] → z = −1
Verify: 2(2) + 3 − (−1) = 4 + 3 + 1 = 8 ✓. Enter this matrix above to see all 7 row operations in detail.
Worked Example: 2×3 Augmented Matrix (Infinite Solutions)
Solve: x + 2y + z = 4, 2x + 4y + 2z = 8.
[ 1 2 1 | 4 ]
[ 2 4 2 | 8 ] R₂ → R₂ − 2R₁ → [ 0 0 0 | 0 ]
RREF: pivot in column 1 only. Columns 2 (y) and 3 (z) are free. The solution is x = 4 − 2s − t, y = s, z = t for any s, t ∈ ℝ. This is a 2-dimensional solution set (a plane in ℝ³ passing through the origin after shifting).
Converting Word Problems to Augmented Matrices
Any linear relationship can be expressed as a row in an augmented matrix. The coefficients of the unknowns fill the left side; the constant goes on the right.
Example: A shop sells pens for $2 and notebooks for $5. A customer buys 3 pens and 2 notebooks for $16. Another buys 1 pen and 4 notebooks for $22. How many of each were sold to a third customer who spent $14?
Variables: p = pens, n = notebooks. The known transactions give:
3p + 2n = 16
1p + 4n = 22
Augmented matrix: [[3, 2, | 16], [1, 4, | 22]]. RREF gives p = 2, n = 5. The unit prices are $2 and $5 — verifiable as consistent.
Augmented vs. Coefficient Matrix: Key Differences
The coefficient matrix A contains only variable coefficients — it encodes the structure of the transformation Ax. Its RREF tells you: the rank (number of pivots), whether A is invertible (rank = n), and the null space (free variables in Ax = 0).
The augmented matrix [A|b] encodes the specific system Ax = b. Its RREF tells you the solution set for that particular b. Changing b changes the solution but not A — this is why Gaussian elimination on [A|b] is more efficient than computing A⁻¹ when you only need one solution.
When you need the matrix inverse A⁻¹ (which handles all right-hand sides simultaneously), apply Gauss-Jordan to [A|I] instead of [A|b]. When you only need one particular solution to Ax = b, the augmented matrix approach is more efficient.
Augmented Matrices and the Rank-Nullity Theorem
The Rouché–Capelli theorem states: the system Ax = b is consistent if and only if rank(A) = rank([A|b]). If consistent, the solution set has dimension n − rank(A) (where n is the number of variables). This dimension is the number of free parameters.
For a 4×4 system where rank(A) = 3: one free variable, so infinite solutions form a line in ℝ⁴. For rank(A) = 2: two free variables, solutions form a plane. For rank(A) = 0: A is the zero matrix, and if b ≠ 0 there is no solution; if b = 0, all of ℝⁿ is the solution.
Frequently Asked Questions
What is an augmented matrix?
An augmented matrix [A|b] combines the coefficient matrix A with the constant vector b. Each row encodes one equation of the linear system Ax = b.
How do I read the solution from RREF of [A|b]?
Each pivot row: the variable in the pivot column equals the value in the last column. Non-pivot columns are free variables — assign them parameters t₁, t₂, … and express everything else in terms of them.
What does a zero row mean in the augmented RREF?
A zero row in the coefficient part with zero in the constant column: redundant equation (no information). With non-zero constant: the system is inconsistent (no solution).
Can I enter more equations than unknowns?
Yes. An m×n augmented matrix for m equations and n unknowns is handled the same way. Extra equations either add no information (redundant) or make the system inconsistent.
What is the difference between augmented and coefficient matrices?
The coefficient matrix A is used for structural properties (rank, invertibility, null space). The augmented matrix [A|b] is used to solve a specific system Ax = b. RREF of A ignores b; RREF of [A|b] solves for x given b.