Skip to content

Elimination Method Calculator

Solve 2×2 and 3×3 systems of equations by elimination with every row operation shown step by step.

Enter the coefficients of each equation, with their own sign — type -1 for a term like −y. The calculator runs Gaussian elimination on the augmented matrix and shows every row operation. Accepts integers, decimals, or fractions like 1/2.
1. x + y =
2. x + y =

Results

Solution

Quick Answer

This elimination method calculator solves a 2×2 or 3×3 system of linear equations by eliminating variables one at a time: multiply one equation by a constant, subtract it from another to remove a variable, and repeat until one unknown remains, then back-substitute. It runs Gaussian elimination on the augmented matrix, shows every row operation, and honestly reports unique, infinitely many, or no solutions.

How the Elimination Method Calculator Works

The elimination method removes one variable at a time by adding a multiple of one equation to another. This calculator does exactly that, but it keeps the bookkeeping in an augmented matrix — the coefficients on the left, the right-hand sides after the bar — so every step is one visible row operation instead of a rewritten pair of equations.

The one operation that does all the work:
Ri → Ri − (factor) × Rk, where factor = aik ÷ akk. Choosing that factor is what makes the leading coefficient cancel to zero.

For the two-equation system 2x + 3y = 13 and 4x − y = 5, the augmented matrix starts as [2 3 | 13 ; 4 −1 | 5]. One row operation clears the second row’s x term, back-substitution recovers both variables, and the calculator prints x = 2, y = 3.

Row operationWhat it doesDoes it change the solution?
Swap two rowsReorders the equationsNo
Multiply a row by a non-zero numberScales one equationNo
Add a multiple of one row to anotherEliminates a variableNo

All three are reversible, which is the reason the method is trustworthy: the system you finish with has the same solution set as the one you started with. Wolfram MathWorld’s entry on Gaussian elimination sets out the same three operations formally.

Why the Calculator Reorders Your Equations

Enter a system and the first line of the working will often read Swap R1 ↔ R2 (partial pivoting). That is not an error, and it is the single most confusing thing this calculator does — so it is worth stating plainly what triggers it.

Before clearing a column, the calculator looks down that column and moves the row with the largest absolute value to the top. Dividing by the largest available number keeps the multiplying factors small, which stops rounding error from compounding down the matrix. Wikipedia’s article on the pivot element covers the same rule.

Running every non-singular two-variable system with whole-number coefficients from −3 to 3 — 2,112 of them — the calculator swaps rows before eliminating in 804 cases, 38.1%. Reordering is the normal case, not the exception, so a step list that opens with a swap is working correctly.

The calculator’s own opening example is one of them. It shows 2x + 3y = 13 first, but 4 is larger than 2, so the working begins by putting the second equation on top:

StepOperationMatrix
Start[2 3 | 13 ; 4 −1 | 5]
1Swap R1 ↔ R2 (partial pivoting)[4 −1 | 5 ; 2 3 | 13]
2R2 → R2 − (0.5) × R1[4 −1 | 5 ; 0 3.5 | 10.5]
Back-substitute3.5y = 10.5, then 4x − 3 = 5x = 2, y = 3

If you are checking the calculator against homework you solved by hand, the row order will frequently differ and the answer will not. A hand solution that never swaps rows is not wrong — it is just skipping a safeguard that only matters once the arithmetic stops being exact.

Unique, None, or Infinitely Many

Every linear system ends in exactly one of three states, and the final matrix tells you which before the calculator says a word. Watch the bottom row.

Bottom row of the reduced matrixReads asResult
A non-zero leading coefficientA real equation in one variableUnique solution
[0 0 | non-zero]0 = something that is not 0No solution
[0 0 | 0]0 = 0, no informationInfinitely many solutions

The pair below is the reason this matters. Both systems have determinant 0, so neither has a unique solution — but they are not the same failure, and a determinant on its own cannot tell them apart.

SystemReduces toDeterminantCalculator says
2x + 3y = 6 and 4x + 6y = 15[4 6 | 15 ; 0 0 | −1.5]0No solution
2x + 3y = 6 and 4x + 6y = 12[4 6 | 12 ; 0 0 | 0]0Infinitely many solutions

Geometrically the first pair is two parallel lines that never meet; the second is the same line written twice. The coefficients are identical in both — only the right-hand side moves, from 15 to 12 — which is why the constant column has to travel through the elimination with the coefficients rather than being handled separately.

How to Use the Elimination Method Calculator

  1. Choose 2 × 2 or 3 × 3.

    Two equations in x and y, or three in x, y and z. The grid redraws to match.

  2. Type each coefficient with its own sign.

    Every connector on screen is a plus, so a term like −y is entered as -1, not as 1. This is the one thing to get right: the calculator solves the numbers you type, not the equation you meant.

  3. Use fractions if that is what the question gives you.

    Entries such as 1/2 or -3/4 are accepted directly, so there is no need to convert to decimals first and introduce rounding you did not have.

  4. Read the solution, then read the working.

    The result card names the outcome — unique, none, or infinitely many. Below it, the augmented matrix is shown before and after, with every row operation in between.

  5. Check the swap line before comparing with your own work.

    If step 1 is a pivot swap, your rows and the calculator’s rows are in different orders from that point on. Compare the final answer, not the intermediate matrices.

Worked Examples

SystemRow operationsResult
2x + 3y = 13 ; 4x − y = 52x = 2, y = 3
x + 5y = 6 ; 2x + 3y = 52x = 1, y = 1
3x + y = 5 ; x + 2y = 41x = 1.2, y = 1.4
x + y = 1 ; x − 2y = 01x = 0.6667, y = 0.3333
2x + 3y = 6 ; 4x + 6y = 152No solution
2x + 3y = 6 ; 4x + 6y = 122Infinitely many solutions
x + y + z = 6 ; 2x − y + 3z = 13 ; −x + 2y + 2z = 94x = 1, y = 1, z = 4

The third and fourth rows are the pair worth studying together. In 3x + y = 5 ; x + 2y = 4 the calculator prints a factor of 0.3333 and an intermediate row of [0 1.6667 | 2.3333] — visibly rounded — and still returns x = 1.2, y = 1.4, which is exact. In x + y = 1 ; x − 2y = 0 the factor is a clean 1 and the intermediate row is the exact [0 −3 | −1], yet the answer prints as 0.6667 and 0.3333, because it genuinely is two thirds and one third.

Rounded working does not imply a rounded answer, and an exact-looking matrix does not imply an exact answer. The two are independent, because the calculator carries full precision internally and rounds only when it draws.

What the Calculator Rounds, and What It Does Not

Display is rounded to four decimal places; the arithmetic is not. Every number on screen — matrix entries, elimination factors, the solution itself — passes through a four-decimal rounding on its way to the page. The values feeding the next step are the full-precision ones.

Fractional answers are common, not rare. Across the same 2,112 systems, the exact solution is a non-terminating decimal in 808 cases, 38.3%. When you see 0.3333 or 7.6667, the honest answer is a fraction — one third and twenty-three thirds in those two cases — and a homework answer is usually expected in that form. The calculator does not convert back, so read a repeating tail as a signal to work out the fraction yourself.

It solves square systems only. Two equations in two unknowns, or three in three. A system with more equations than unknowns, or fewer, needs a different tool.

It will not tell you which variable to eliminate first. That decision is made for you by partial pivoting, which optimises for numerical stability rather than for tidy arithmetic. By hand you would usually pick the variable with a coefficient of 1; the calculator picks the largest.

Elimination applies to linear equations only — every variable to the first power, none multiplied together. A single equation with a squared term needs the quadratic formula calculator instead. And a pair of conditions like “two numbers with sum 7 and product 12” looks like a two-variable system but is not linear, because the product multiplies the unknowns together; that one is the diamond problem calculator.

Elimination Method Calculator: FAQ

It uses partial pivoting: before clearing a column it moves the row with the largest absolute leading value to the top, which keeps the multiplying factors small and limits rounding error. Across 2,112 two-variable systems with coefficients from −3 to 3, it swaps in 804 of them — 38.1%. Row swaps never change the solution set, so your hand working and the calculator’s can differ step by step and still agree on the answer.

Type the minus sign into the coefficient box itself. Every connector between terms on screen is a plus, so −y is entered as -1 and −2.5z as -2.5. The calculator solves exactly the numbers in the boxes, so a term entered without its sign will quietly give you the answer to a different system.

Look at the bottom row after elimination. A row reading [0 0 | non-zero] says 0 equals something that is not zero, which is a contradiction — no solution, two parallel lines. A row reading [0 0 | 0] says 0 = 0, which is true but tells you nothing — infinitely many solutions, the same line twice. Both have determinant zero, so the determinant alone cannot separate them.

The calculator prints decimals rounded to four places and does not convert back to fractions. A repeating tail such as 0.6667, 0.3333 or 7.6667 means the exact answer is a fraction — two thirds, one third and twenty-three thirds respectively. This is not unusual: over 2,112 test systems the exact answer was a non-terminating decimal 38.3% of the time. If your homework wants a fraction, convert the decimal back yourself.

Yes. Entries like 1/2, -3/4 and 7/8 are parsed directly, as are ordinary decimals. Entering the fraction is better than converting it first, because typing 0.667 for two thirds introduces an error the calculator then carries through every row operation.

They are the same idea at two levels of formality. The elimination method taught in algebra adds a multiple of one equation to another to cancel a variable. Gaussian elimination does that systematically on an augmented matrix, clearing one column at a time into row echelon form, then back-substitutes. This calculator runs the matrix version and prints each row operation, so the working matches what you would do by hand apart from the pivot swaps.

No. Every variable must appear to the first power and none may be multiplied by another. x + y = 7 with xy = 12 is not a linear system, because the second condition multiplies the unknowns together, and no sequence of row operations will separate them. A single equation containing a squared term needs the quadratic formula instead.

Three eliminations clear the lower triangle — two in the first column, one in the second — plus any pivot swaps. The calculator’s own 3 × 3 example, x + y + z = 6 with 2x − y + 3z = 13 and −x + 2y + 2z = 9, takes four listed steps and returns x = 1, y = 1, z = 4. A system needing more than that usually has a zero where a pivot was expected.
Made with care

Want a calculator
built just for you?

Share the formula, the inputs, and who it's for — most reader-requested calculators ship within 14 days.

Calculator illustration