Inverse Matrix Method Calculator

Solve a linear system Ax = b by computing A⁻¹ using Gauss-Jordan, then multiplying x = A⁻¹b. All steps shown in full.

100% Free No Sign-Up Calculated in Your Browser

System Size

Coefficient Matrix A

Constants b

← Solve Ax = b
using x = A⁻¹b

Solve a system of linear equations Ax = b by first computing the inverse of A using Gauss-Jordan elimination, then multiplying x = A⁻¹b, with every step of both phases shown.

How It Works

How Inverse Matrix Method Calculator Works

Phase 1 builds an augmented matrix [A | I] by placing the identity matrix next to your coefficient matrix A, then row-reduces the whole thing using Gauss-Jordan elimination (with partial pivoting for stability) until the left half becomes the identity matrix. Whatever the right half has become at that point is A⁻¹.

Phase 2 multiplies A⁻¹ by your constants vector b to get the solution vector x = A⁻¹b — each unknown xᵢ is computed as the dot product of row i of A⁻¹ with b, and every unknown is shown individually before being combined into the final solution.

If A turns out to be singular (no valid pivot can be found in some column, meaning its determinant is zero), the method reports that no inverse — and therefore no unique solution — exists. Otherwise, a final verification step plugs the solution back into Ax to confirm it reproduces b.

Worked Example

See It In Action

For the system 2x + y = 4 and 5x + 3y = 7 (A = [[2,1],[5,3]], b = [4,7]): the determinant of A is (2×3 − 1×5) = 1, so A⁻¹ = [[3,-1],[-5,2]]. Multiplying x = A⁻¹b gives x = 3(4) + (−1)(7) = 5, and y = −5(4) + 2(7) = −6. Checking: 2(5) + (−6) = 4 ✓ and 5(5) + 3(−6) = 25 − 18 = 7 ✓ — both original equations balance.

Real-World Use Cases

Who Uses Inverse Matrix Method Calculator and Why

Solving Ax = b when the same coefficient matrix A will be reused against several different b vectors.
Learning the two-phase technique of first computing A⁻¹, then multiplying it by b to get the solution.
Verifying a solved system by checking that the final Ax reproduces the original b vector.
Comparing results against Cramer's Rule or Gauss-Jordan elimination for the same system as an independent check.

Common Mistakes

Mistakes to Avoid

Applying this method to a system where A is singular — if no valid pivot exists during elimination (determinant of zero), no inverse exists and this method can't produce a solution.
Skipping the final Ax = b verification step, which is the clearest way to catch an arithmetic mistake made in either the inversion phase or the multiplication phase.
Assuming this method is functionally identical to Cramer's Rule — this approach builds the full A⁻¹ first and then multiplies by b, while Cramer's Rule solves for each variable directly via determinant ratios; both fail when A is singular, but by different routes.

Pro Tips

Tips for Best Results

Reuse the computed A⁻¹ for additional constants vectors without repeating the elimination phase — that's the main advantage of this method over row-reducing from scratch each time.
Always run the final verification (plugging the solution back into Ax) to confirm no arithmetic slip occurred in either phase.

Troubleshooting

Fixing Common Problems

The calculator reports no inverse exists. — This means A is singular (no valid pivot could be found during elimination) — the system either has no solution or infinitely many, and this method can't distinguish which; try row-reducing the augmented system directly instead.

My solution doesn't satisfy the original equations. — Recheck the inversion phase first, since an error there carries through to every subsequent step of the multiplication phase.

Glossary

Terms Explained

Augmented matrix: A matrix formed by placing the identity matrix next to A, used to compute A⁻¹ via Gauss-Jordan elimination.

Verification step: Substituting the computed solution back into the original Ax = b equations to confirm it reproduces b exactly.

FAQ

Frequently Asked Questions

Why compute the full inverse instead of just solving directly?

Computing A⁻¹ explicitly is useful when you need to solve Ax = b for several different b vectors using the same A — once A⁻¹ is known, each new solution is just one matrix-vector multiplication away, rather than re-running elimination from scratch.

What does it mean if the system has no unique solution?

If A is singular (its determinant is zero), no inverse exists, which means the system either has no solution at all or has infinitely many — the inverse matrix method cannot resolve which, so a different technique (like row-reducing the augmented system directly) is needed to tell them apart.

Is the inverse matrix method the same as Cramer's rule?

They're related but distinct — Cramer's rule solves for each variable directly using ratios of determinants, while the inverse matrix method first builds the full A⁻¹ and then multiplies it by b. Both fail when A is singular, but they get there by different routes.

Does this work for systems larger than 3×3?

This calculator supports 2×2 and 3×3 systems. The same Gauss-Jordan approach extends to larger systems in principle, but tracking every step by hand (or on screen) becomes unwieldy well before you reach much larger sizes.