diff --git a/cpp/src/mip/problem/problem_helpers.cuh b/cpp/src/mip/problem/problem_helpers.cuh
index 2104f7eaf..ddb9cbe7e 100644
--- a/cpp/src/mip/problem/problem_helpers.cuh
+++ b/cpp/src/mip/problem/problem_helpers.cuh
@@ -138,7 +138,6 @@ static void convert_to_maximization_problem(detail::problem_t<i_t, f_t>& op_prob
   // negating objective coeffs
   op_problem.presolve_data.objective_scaling_factor =
     -op_problem.presolve_data.objective_scaling_factor;
-  op_problem.presolve_data.objective_offset = -op_problem.presolve_data.objective_offset;
 }
 
 /*
diff --git a/cpp/tests/linear_programming/pdlp_test.cu b/cpp/tests/linear_programming/pdlp_test.cu
index 212e4e1d3..a9e042812 100644
--- a/cpp/tests/linear_programming/pdlp_test.cu
+++ b/cpp/tests/linear_programming/pdlp_test.cu
@@ -906,6 +906,42 @@ TEST(pdlp_class, run_empty_matrix_dual_simplex)
   EXPECT_FALSE(solution.get_additional_termination_information().solved_by_pdlp);
 }
 
+TEST(pdlp_class, test_max)
+{
+  const raft::handle_t handle_{};
+
+  auto path = make_path_absolute("linear_programming/good-max.mps");
+  cuopt::mps_parser::mps_data_model_t<int, double> op_problem =
+    cuopt::mps_parser::parse_mps<int, double>(path);
+
+  auto solver_settings   = pdlp_solver_settings_t<int, double>{};
+  solver_settings.method = cuopt::linear_programming::method_t::PDLP;
+
+  optimization_problem_solution_t<int, double> solution =
+    solve_lp(&handle_, op_problem, solver_settings);
+  EXPECT_EQ((int)solution.get_termination_status(), CUOPT_TERIMINATION_STATUS_OPTIMAL);
+  EXPECT_NEAR(
+    solution.get_additional_termination_information().primal_objective, 17.0, factor_tolerance);
+}
+
+TEST(pdlp_class, test_max_with_offset)
+{
+  const raft::handle_t handle_{};
+
+  auto path = make_path_absolute("linear_programming/max_offset.mps");
+  cuopt::mps_parser::mps_data_model_t<int, double> op_problem =
+    cuopt::mps_parser::parse_mps<int, double>(path);
+
+  auto solver_settings   = pdlp_solver_settings_t<int, double>{};
+  solver_settings.method = cuopt::linear_programming::method_t::PDLP;
+
+  optimization_problem_solution_t<int, double> solution =
+    solve_lp(&handle_, op_problem, solver_settings);
+  EXPECT_EQ((int)solution.get_termination_status(), CUOPT_TERIMINATION_STATUS_OPTIMAL);
+  EXPECT_NEAR(
+    solution.get_additional_termination_information().primal_objective, 0.0, factor_tolerance);
+}
+
 }  // namespace cuopt::linear_programming::test
 
 CUOPT_TEST_PROGRAM_MAIN()
diff --git a/datasets/linear_programming/good-max.mps b/datasets/linear_programming/good-max.mps
new file mode 100644
index 000000000..dbf3249d7
--- /dev/null
+++ b/datasets/linear_programming/good-max.mps
@@ -0,0 +1,19 @@
+NAME          MAXEX
+OBJSENSE
+ MAX
+ROWS
+ N  OBJ
+ L  LIM1
+COLUMNS
+    X1    OBJ     2    LIM1    1
+    X2    OBJ     3    LIM1    2
+    X3    OBJ     1    LIM1    1
+RHS
+    RHS1  LIM1    10
+BOUNDS
+ LO BND1  X1      0
+ UP BND1  X1      4
+ LO BND1  X2      0
+ UP BND1  X2      6
+ LO BND1  X3      0
+ENDATA
\ No newline at end of file
diff --git a/datasets/linear_programming/max_offset.mps b/datasets/linear_programming/max_offset.mps
new file mode 100644
index 000000000..99ee7a69f
--- /dev/null
+++ b/datasets/linear_programming/max_offset.mps
@@ -0,0 +1,25 @@
+NAME
+OBJSENSE
+  MAXIMIZE
+ROWS
+ N  OBJ
+ L  c1
+COLUMNS
+    x1        c1        3
+    x1        OBJ       1
+    x2        c1        2
+    x2        OBJ       2
+    x3        c1        1
+    x3        OBJ       3
+RHS
+    rhs       c1        2
+    rhs       OBJ       4
+RANGES
+BOUNDS
+ LO bounds    x1        0
+ PL bounds    x1
+ LO bounds    x2        0
+ PL bounds    x2
+ LO bounds    x3        0
+ UP bounds    x3        1
+ENDATA
\ No newline at end of file