diff --git a/src/SolverTest.jl b/src/SolverTest.jl
index 6abaadf..f1f0531 100644
--- a/src/SolverTest.jl
+++ b/src/SolverTest.jl
@@ -10,6 +10,7 @@ include("kkt.jl")
 include("nlp/unconstrained.jl")
 include("nlp/bound-constrained.jl")
 include("nlp/equality-constrained.jl")
+include("nlp/inequality-constrained.jl")
 include("nlp/multiprecision.jl")
 
 include("nls/unconstrained.jl")
diff --git a/src/nlp/inequality-constrained.jl b/src/nlp/inequality-constrained.jl
new file mode 100644
index 0000000..477db88
--- /dev/null
+++ b/src/nlp/inequality-constrained.jl
@@ -0,0 +1,81 @@
+export inequality_constrained_nlp
+
+function inequality_constrained_nlp_set()
+  return [
+    ADNLPModel( # ones(2),
+      x -> 2x[1]^2 + x[1] * x[2] + x[2]^2 - 9x[1] - 9x[2],
+      [1.0; 2.0],
+      x -> [4x[1] + 6x[2] - 10],
+      [-Inf],
+      [0.0],
+      name = "Simple quadratic problem with a linear constraint",
+    ),
+    ADNLPModel( # [0.990099; 0.980296],
+      x -> (x[1] - 1)^2 + 0.01 * x[2],
+      [-1.2; 1.0],
+      x -> [10 * (x[2] - x[1]^2)],
+      [0.0],
+      [Inf],
+      name = "Simple quadratic problem with a quadratic constraint",
+    ),
+    ADNLPModel(
+      x -> x[1] - x[2],
+      [-10.0; 10.0],
+      x -> [-3 * x[1]^2 + 2 * x[1] * x[2] - x[2]^2 + 1],
+      [0.0],
+      [Inf],
+      name = "HS10",
+    ),
+    ADNLPModel( # ones(2),
+      x -> (x[1] - 1)^2 + 100 * (x[2] - x[1]^2)^2,
+      [-1.2; 1.0],
+      x -> [x[1]^2 + x[2]^2],
+      [0.0],
+      [4.0],
+      name = "Rosenbrock with circle constraint",
+    ),
+    ADNLPModel( # [5.196152; 1.732051],
+      x -> -x[1],
+      [2.0; 1.0],
+      x -> [x[1]^2 + x[2]^2; x[1] * x[2]],
+      [25.0; 9.0],
+      [30.0; 12.0],
+      name = "Unbounded objective with inequalities",
+    ),
+    ADNLPModel(
+      x -> (x[1] - 2)^2 + (x[2] - 1)^2,
+      [-10.0; 10.0],
+      x -> [x[1] + x[2], -x[1]^2 + x[2]],
+      [-Inf; 0.0],
+      [2; Inf],
+      name = "HS22",
+    ),
+  ]
+end
+
+"""
+    inequality_constrained_nlp(solver; problem_set = inequality_constrained_nlp_set(), atol = 1e-6, rtol = 1e-6)
+
+Test the `solver` on inequality-constrained problems.
+If `rtol` is non-zero, the relative error uses the gradient at the initial guess.
+"""
+function inequality_constrained_nlp(
+  solver;
+  problem_set = inequality_constrained_nlp_set(),
+  atol = 1e-6,
+  rtol = 1e-6,
+)
+  @testset "Problem $(nlp.meta.name)" for nlp in problem_set
+    stats = with_logger(NullLogger()) do
+      solver(nlp)
+    end
+    ng0 = rtol != 0 ? norm(grad(nlp, nlp.meta.x0)) : 0
+    ϵ = atol + rtol * ng0
+    primal, dual = kkt_checker(nlp, stats.solution)
+    @test all(abs.(dual) .< ϵ)
+    @test all(abs.(primal) .< ϵ)
+    @test stats.dual_feas < ϵ
+    @test stats.primal_feas < ϵ
+    @test stats.status == :first_order
+  end
+end