diff --git a/ortools/sat/python/linear_expr.cc b/ortools/sat/python/linear_expr.cc
index fe6e7c6787..df649f8a9b 100644
--- a/ortools/sat/python/linear_expr.cc
+++ b/ortools/sat/python/linear_expr.cc
@@ -189,8 +189,8 @@ LinearExpr* LinearExpr::Constant(int64_t value) {
   return new IntConstant(value);
 }
 
-LinearExpr* LinearExpr::Add(LinearExpr* other) {
-  return new BinaryAdd(this, other);
+LinearExpr* LinearExpr::Add(LinearExpr* expr) {
+  return new BinaryAdd(this, expr);
 }
 
 LinearExpr* LinearExpr::AddInt(int64_t cst) {
@@ -203,40 +203,40 @@ LinearExpr* LinearExpr::AddDouble(double cst) {
   return new FloatAffine(this, 1.0, cst);
 }
 
-LinearExpr* LinearExpr::Sub(ExprOrValue other) {
-  if (other.expr != nullptr) {
-    return new IntWeightedSum({this, other.expr}, {1, -1}, 0);
-  } else if (other.double_value != 0.0) {
-    return new FloatAffine(this, 1.0, -other.double_value);
-  } else if (other.int_value != 0) {
-    return new IntAffine(this, 1, -other.int_value);
-  } else {
-    return this;
-  }
+LinearExpr* LinearExpr::Sub(LinearExpr* expr) {
+  return new IntWeightedSum({this, expr}, {1, -1}, 0);
 }
 
-LinearExpr* LinearExpr::RSub(ExprOrValue other) {
-  if (other.expr != nullptr) {
-    return new IntWeightedSum({this, other.expr}, {-1, 1}, 0);
-  } else if (other.double_value != 0.0) {
-    return new FloatAffine(this, -1.0, other.double_value);
-  } else {
-    return new IntAffine(this, -1, other.int_value);
-  }
+LinearExpr* LinearExpr::SubInt(int64_t cst) {
+  if (cst == 0) return this;
+  return new IntAffine(this, 1, -cst);
 }
 
-LinearExpr* LinearExpr::Mul(double cst) {
-  if (cst == 0.0) return new IntConstant(0);
-  if (cst == 1.0) return this;
-  return new FloatAffine(this, cst, 0.0);
+LinearExpr* LinearExpr::SubDouble(double cst) {
+  if (cst == 0.0) return this;
+  return new FloatAffine(this, 1.0, -cst);
 }
 
-LinearExpr* LinearExpr::Mul(int64_t cst) {
+LinearExpr* LinearExpr::RSubInt(int64_t cst) {
+  return new IntAffine(this, -1, cst);
+}
+
+LinearExpr* LinearExpr::RSubDouble(double cst) {
+  return new FloatAffine(this, -1.0, cst);
+}
+
+LinearExpr* LinearExpr::MulInt(int64_t cst) {
   if (cst == 0) return new IntConstant(0);
   if (cst == 1) return this;
   return new IntAffine(this, cst, 0);
 }
 
+LinearExpr* LinearExpr::MulDouble(double cst) {
+  if (cst == 0.0) return new IntConstant(0);
+  if (cst == 1.0) return this;
+  return new FloatAffine(this, cst, 0.0);
+}
+
 LinearExpr* LinearExpr::Neg() { return new IntAffine(this, -1, 0); }
 
 void FloatExprVisitor::AddToProcess(LinearExpr* expr, double coeff) {
@@ -449,152 +449,142 @@ std::string FloatAffine::DebugString() const {
   return absl::StrCat("FloatAffine(expr=", expr_->DebugString(),
                       ", coeff=", coeff_, ", offset=", offset_, ")");
 }
-
-BoundedLinearExpression* LinearExpr::Eq(ExprOrValue other) {
-  if (other.double_value != 0.0) return nullptr;
-  if (other.expr != nullptr) {
-    IntExprVisitor lin;
-    lin.AddToProcess(this, 1);
-    lin.AddToProcess(other.expr, -1);
-    std::vector<BaseIntVar*> vars;
-    std::vector<int64_t> coeffs;
-    int64_t offset;
-    if (!lin.Process(&vars, &coeffs, &offset)) return nullptr;
-    return new BoundedLinearExpression(vars, coeffs, offset, Domain(0));
-  } else {
-    IntExprVisitor lin;
-    lin.AddToProcess(this, 1);
-    std::vector<BaseIntVar*> vars;
-    std::vector<int64_t> coeffs;
-    int64_t offset;
-    if (!lin.Process(&vars, &coeffs, &offset)) return nullptr;
-    return new BoundedLinearExpression(vars, coeffs, offset,
-                                       Domain(other.int_value));
-  }
+BoundedLinearExpression* LinearExpr::Eq(LinearExpr* rhs) {
+  IntExprVisitor lin;
+  lin.AddToProcess(this, 1);
+  lin.AddToProcess(rhs, -1);
+  std::vector<BaseIntVar*> vars;
+  std::vector<int64_t> coeffs;
+  int64_t offset;
+  if (!lin.Process(&vars, &coeffs, &offset)) return nullptr;
+  return new BoundedLinearExpression(vars, coeffs, offset, Domain(0));
 }
 
-BoundedLinearExpression* LinearExpr::Ne(ExprOrValue other) {
-  if (other.double_value != 0.0) return nullptr;
-  if (other.expr != nullptr) {
-    IntExprVisitor lin;
-    lin.AddToProcess(this, 1);
-    lin.AddToProcess(other.expr, -1);
-    std::vector<BaseIntVar*> vars;
-    std::vector<int64_t> coeffs;
-    int64_t offset;
-    if (!lin.Process(&vars, &coeffs, &offset)) return nullptr;
-    return new BoundedLinearExpression(vars, coeffs, offset,
-                                       Domain(0).Complement());
-  } else {
-    IntExprVisitor lin;
-    lin.AddToProcess(this, 1);
-    std::vector<BaseIntVar*> vars;
-    std::vector<int64_t> coeffs;
-    int64_t offset;
-    if (!lin.Process(&vars, &coeffs, &offset)) return nullptr;
-    return new BoundedLinearExpression(vars, coeffs, offset,
-                                       Domain(other.int_value).Complement());
-  }
+BoundedLinearExpression* LinearExpr::EqCst(int64_t rhs) {
+  IntExprVisitor lin;
+  lin.AddToProcess(this, 1);
+  std::vector<BaseIntVar*> vars;
+  std::vector<int64_t> coeffs;
+  int64_t offset;
+  if (!lin.Process(&vars, &coeffs, &offset)) return nullptr;
+  return new BoundedLinearExpression(vars, coeffs, offset, Domain(rhs));
 }
 
-BoundedLinearExpression* LinearExpr::Le(ExprOrValue other) {
-  if (other.double_value != 0.0) return nullptr;
-  if (other.expr != nullptr) {
-    IntExprVisitor lin;
-    lin.AddToProcess(this, 1);
-    lin.AddToProcess(other.expr, -1);
-    std::vector<BaseIntVar*> vars;
-    std::vector<int64_t> coeffs;
-    int64_t offset;
-    if (!lin.Process(&vars, &coeffs, &offset)) return nullptr;
-    return new BoundedLinearExpression(
-        vars, coeffs, offset, Domain(std::numeric_limits<int64_t>::min(), 0));
-  } else {
-    IntExprVisitor lin;
-    lin.AddToProcess(this, 1);
-    std::vector<BaseIntVar*> vars;
-    std::vector<int64_t> coeffs;
-    int64_t offset;
-    if (!lin.Process(&vars, &coeffs, &offset)) return nullptr;
-    return new BoundedLinearExpression(
-        vars, coeffs, offset,
-        Domain(std::numeric_limits<int64_t>::min(), other.int_value));
-  }
+BoundedLinearExpression* LinearExpr::Ne(LinearExpr* rhs) {
+  IntExprVisitor lin;
+  lin.AddToProcess(this, 1);
+  lin.AddToProcess(rhs, -1);
+  std::vector<BaseIntVar*> vars;
+  std::vector<int64_t> coeffs;
+  int64_t offset;
+  if (!lin.Process(&vars, &coeffs, &offset)) return nullptr;
+  return new BoundedLinearExpression(vars, coeffs, offset,
+                                     Domain(0).Complement());
 }
 
-BoundedLinearExpression* LinearExpr::Lt(ExprOrValue other) {
-  if (other.double_value != 0.0) return nullptr;
-  if (other.expr != nullptr) {
-    IntExprVisitor lin;
-    lin.AddToProcess(this, 1);
-    lin.AddToProcess(other.expr, -1);
-    std::vector<BaseIntVar*> vars;
-    std::vector<int64_t> coeffs;
-    int64_t offset;
-    if (!lin.Process(&vars, &coeffs, &offset)) return nullptr;
-    return new BoundedLinearExpression(
-        vars, coeffs, offset, Domain(std::numeric_limits<int64_t>::min(), -1));
-  } else {
-    IntExprVisitor lin;
-    lin.AddToProcess(this, 1);
-    std::vector<BaseIntVar*> vars;
-    std::vector<int64_t> coeffs;
-    int64_t offset;
-    if (!lin.Process(&vars, &coeffs, &offset)) return nullptr;
-    return new BoundedLinearExpression(
-        vars, coeffs, offset,
-        Domain(std::numeric_limits<int64_t>::min(), other.int_value - 1));
-  }
+BoundedLinearExpression* LinearExpr::NeCst(int64_t rhs) {
+  IntExprVisitor lin;
+  lin.AddToProcess(this, 1);
+  std::vector<BaseIntVar*> vars;
+  std::vector<int64_t> coeffs;
+  int64_t offset;
+  if (!lin.Process(&vars, &coeffs, &offset)) return nullptr;
+  return new BoundedLinearExpression(vars, coeffs, offset,
+                                     Domain(rhs).Complement());
 }
 
-BoundedLinearExpression* LinearExpr::Ge(ExprOrValue other) {
-  if (other.double_value != 0.0) return nullptr;
-  if (other.expr != nullptr) {
-    IntExprVisitor lin;
-    lin.AddToProcess(this, 1);
-    lin.AddToProcess(other.expr, -1);
-    std::vector<BaseIntVar*> vars;
-    std::vector<int64_t> coeffs;
-    int64_t offset;
-    if (!lin.Process(&vars, &coeffs, &offset)) return nullptr;
-    return new BoundedLinearExpression(
-        vars, coeffs, offset, Domain(0, std::numeric_limits<int64_t>::max()));
-  } else {
-    IntExprVisitor lin;
-    lin.AddToProcess(this, 1);
-    std::vector<BaseIntVar*> vars;
-    std::vector<int64_t> coeffs;
-    int64_t offset;
-    if (!lin.Process(&vars, &coeffs, &offset)) return nullptr;
-    return new BoundedLinearExpression(
-        vars, coeffs, offset,
-        Domain(other.int_value, std::numeric_limits<int64_t>::max()));
-  }
+BoundedLinearExpression* LinearExpr::Le(LinearExpr* rhs) {
+  IntExprVisitor lin;
+  lin.AddToProcess(this, 1);
+  lin.AddToProcess(rhs, -1);
+  std::vector<BaseIntVar*> vars;
+  std::vector<int64_t> coeffs;
+  int64_t offset;
+  if (!lin.Process(&vars, &coeffs, &offset)) return nullptr;
+  return new BoundedLinearExpression(
+      vars, coeffs, offset, Domain(std::numeric_limits<int64_t>::min(), 0));
 }
 
-BoundedLinearExpression* LinearExpr::Gt(ExprOrValue other) {
-  if (other.double_value != 0.0) return nullptr;
-  if (other.expr != nullptr) {
-    IntExprVisitor lin;
-    lin.AddToProcess(this, 1);
-    lin.AddToProcess(other.expr, -1);
-    std::vector<BaseIntVar*> vars;
-    std::vector<int64_t> coeffs;
-    int64_t offset;
-    if (!lin.Process(&vars, &coeffs, &offset)) return nullptr;
-    return new BoundedLinearExpression(
-        vars, coeffs, offset, Domain(1, std::numeric_limits<int64_t>::max()));
-  } else {
-    IntExprVisitor lin;
-    lin.AddToProcess(this, 1);
-    std::vector<BaseIntVar*> vars;
-    std::vector<int64_t> coeffs;
-    int64_t offset;
-    if (!lin.Process(&vars, &coeffs, &offset)) return nullptr;
-    return new BoundedLinearExpression(
-        vars, coeffs, offset,
-        Domain(other.int_value + 1, std::numeric_limits<int64_t>::max()));
-  }
+BoundedLinearExpression* LinearExpr::LeCst(int64_t rhs) {
+  IntExprVisitor lin;
+  lin.AddToProcess(this, 1);
+  std::vector<BaseIntVar*> vars;
+  std::vector<int64_t> coeffs;
+  int64_t offset;
+  if (!lin.Process(&vars, &coeffs, &offset)) return nullptr;
+  return new BoundedLinearExpression(
+      vars, coeffs, offset, Domain(std::numeric_limits<int64_t>::min(), rhs));
+}
+
+BoundedLinearExpression* LinearExpr::Lt(LinearExpr* rhs) {
+  IntExprVisitor lin;
+  lin.AddToProcess(this, 1);
+  lin.AddToProcess(rhs, -1);
+  std::vector<BaseIntVar*> vars;
+  std::vector<int64_t> coeffs;
+  int64_t offset;
+  if (!lin.Process(&vars, &coeffs, &offset)) return nullptr;
+  return new BoundedLinearExpression(
+      vars, coeffs, offset, Domain(std::numeric_limits<int64_t>::min(), -1));
+}
+
+BoundedLinearExpression* LinearExpr::LtCst(int64_t rhs) {
+  IntExprVisitor lin;
+  lin.AddToProcess(this, 1);
+  std::vector<BaseIntVar*> vars;
+  std::vector<int64_t> coeffs;
+  int64_t offset;
+  if (!lin.Process(&vars, &coeffs, &offset)) return nullptr;
+  return new BoundedLinearExpression(
+      vars, coeffs, offset,
+      Domain(std::numeric_limits<int64_t>::min(), rhs - 1));
+}
+
+BoundedLinearExpression* LinearExpr::Ge(LinearExpr* rhs) {
+  IntExprVisitor lin;
+  lin.AddToProcess(this, 1);
+  lin.AddToProcess(rhs, -1);
+  std::vector<BaseIntVar*> vars;
+  std::vector<int64_t> coeffs;
+  int64_t offset;
+  if (!lin.Process(&vars, &coeffs, &offset)) return nullptr;
+  return new BoundedLinearExpression(
+      vars, coeffs, offset, Domain(0, std::numeric_limits<int64_t>::max()));
+}
+
+BoundedLinearExpression* LinearExpr::GeCst(int64_t rhs) {
+  IntExprVisitor lin;
+  lin.AddToProcess(this, 1);
+  std::vector<BaseIntVar*> vars;
+  std::vector<int64_t> coeffs;
+  int64_t offset;
+  if (!lin.Process(&vars, &coeffs, &offset)) return nullptr;
+  return new BoundedLinearExpression(
+      vars, coeffs, offset, Domain(rhs, std::numeric_limits<int64_t>::max()));
+}
+
+BoundedLinearExpression* LinearExpr::Gt(LinearExpr* rhs) {
+  IntExprVisitor lin;
+  lin.AddToProcess(this, 1);
+  lin.AddToProcess(rhs, -1);
+  std::vector<BaseIntVar*> vars;
+  std::vector<int64_t> coeffs;
+  int64_t offset;
+  if (!lin.Process(&vars, &coeffs, &offset)) return nullptr;
+  return new BoundedLinearExpression(
+      vars, coeffs, offset, Domain(1, std::numeric_limits<int64_t>::max()));
+}
+
+BoundedLinearExpression* LinearExpr::GtCst(int64_t rhs) {
+  IntExprVisitor lin;
+  lin.AddToProcess(this, 1);
+  std::vector<BaseIntVar*> vars;
+  std::vector<int64_t> coeffs;
+  int64_t offset;
+  if (!lin.Process(&vars, &coeffs, &offset)) return nullptr;
+  return new BoundedLinearExpression(
+      vars, coeffs, offset,
+      Domain(rhs + 1, std::numeric_limits<int64_t>::max()));
 }
 
 void IntExprVisitor::AddToProcess(LinearExpr* expr, int64_t coeff) {
diff --git a/ortools/sat/python/linear_expr.h b/ortools/sat/python/linear_expr.h
index 5c38595e65..9ed2a35c42 100644
--- a/ortools/sat/python/linear_expr.h
+++ b/ortools/sat/python/linear_expr.h
@@ -78,21 +78,30 @@ class LinearExpr {
   static LinearExpr* Constant(int64_t value);
   static LinearExpr* Constant(double value);
 
-  LinearExpr* Add(LinearExpr* other);
+  LinearExpr* Add(LinearExpr* expr);
   LinearExpr* AddInt(int64_t cst);
   LinearExpr* AddDouble(double cst);
-  LinearExpr* Sub(ExprOrValue other);
-  LinearExpr* RSub(ExprOrValue other);
-  LinearExpr* Mul(double cst);
-  LinearExpr* Mul(int64_t cst);
+  LinearExpr* Sub(LinearExpr* expr);
+  LinearExpr* SubInt(int64_t cst);
+  LinearExpr* SubDouble(double cst);
+  LinearExpr* RSubInt(int64_t cst);
+  LinearExpr* RSubDouble(double cst);
+  LinearExpr* MulInt(int64_t cst);
+  LinearExpr* MulDouble(double cst);
   LinearExpr* Neg();
 
-  BoundedLinearExpression* Eq(ExprOrValue other);
-  BoundedLinearExpression* Ne(ExprOrValue other);
-  BoundedLinearExpression* Ge(ExprOrValue other);
-  BoundedLinearExpression* Le(ExprOrValue other);
-  BoundedLinearExpression* Lt(ExprOrValue other);
-  BoundedLinearExpression* Gt(ExprOrValue other);
+  BoundedLinearExpression* Eq(LinearExpr* rhs);
+  BoundedLinearExpression* EqCst(int64_t rhs);
+  BoundedLinearExpression* Ne(LinearExpr* rhs);
+  BoundedLinearExpression* NeCst(int64_t rhs);
+  BoundedLinearExpression* Ge(LinearExpr* rhs);
+  BoundedLinearExpression* GeCst(int64_t rhs);
+  BoundedLinearExpression* Le(LinearExpr* rhs);
+  BoundedLinearExpression* LeCst(int64_t rhs);
+  BoundedLinearExpression* Lt(LinearExpr* rhs);
+  BoundedLinearExpression* LtCst(int64_t rhs);
+  BoundedLinearExpression* Gt(LinearExpr* rhs);
+  BoundedLinearExpression* GtCst(int64_t rhs);
 };
 
 // Compare the indices of variables.
diff --git a/ortools/sat/python/swig_helper.cc b/ortools/sat/python/swig_helper.cc
index a6b5e42111..09000f8172 100644
--- a/ortools/sat/python/swig_helper.cc
+++ b/ortools/sat/python/swig_helper.cc
@@ -394,18 +394,18 @@ PYBIND11_MODULE(swig_helper, m) {
                                     const std::vector<double>&>(
                       &LinearExpr::WeightedSum),
                   py::return_value_policy::automatic, py::keep_alive<0, 1>())
-      .def_static("Term",
-                  py::overload_cast<LinearExpr*, int64_t>(&LinearExpr::Term),
-                  arg("expr"), arg("coeff"), "Returns expr * coeff.",
-                  py::return_value_policy::automatic, py::keep_alive<0, 1>())
-      .def_static("Term",
-                  py::overload_cast<LinearExpr*, double>(&LinearExpr::Term),
-                  arg("expr"), arg("coeff"), "Returns expr * coeff.",
-                  py::return_value_policy::automatic, py::keep_alive<0, 1>())
+      .def_static(
+          "Term", py::overload_cast<LinearExpr*, int64_t>(&LinearExpr::Term),
+          arg("expr").none(false), arg("coeff"), "Returns expr * coeff.",
+          py::return_value_policy::automatic, py::keep_alive<0, 1>())
+      .def_static(
+          "Term", py::overload_cast<LinearExpr*, double>(&LinearExpr::Term),
+          arg("expr").none(false), arg("coeff"), "Returns expr * coeff.",
+          py::return_value_policy::automatic, py::keep_alive<0, 1>())
       .def("__str__", &LinearExpr::ToString)
       .def("__repr__", &LinearExpr::DebugString)
       .def("is_integer", &LinearExpr::IsInteger)
-      .def("__add__", &LinearExpr::Add, arg("other"),
+      .def("__add__", &LinearExpr::Add, arg("other").none(false),
            py::return_value_policy::automatic, py::keep_alive<0, 1>(),
            py::keep_alive<0, 2>())
       .def("__add__", &LinearExpr::AddInt, arg("cst"),
@@ -416,34 +416,76 @@ PYBIND11_MODULE(swig_helper, m) {
            py::return_value_policy::automatic, py::keep_alive<0, 1>())
       .def("__radd__", &LinearExpr::AddDouble, arg("cst"),
            py::return_value_policy::automatic, py::keep_alive<0, 1>())
-      .def("__sub__", &LinearExpr::Sub, arg("other"),
+      .def("__sub__", &LinearExpr::Sub, arg("other").none(false),
            py::return_value_policy::automatic, py::keep_alive<0, 1>(),
            py::keep_alive<0, 2>())
-      .def("__rsub__", &LinearExpr::RSub, arg("other"),
-           py::return_value_policy::automatic, py::keep_alive<0, 1>(),
-           py::keep_alive<0, 2>())
-      .def("__mul__", py::overload_cast<double>(&LinearExpr::Mul), arg("cst"),
+      .def("__sub__", &LinearExpr::SubInt, arg("cst"),
            py::return_value_policy::automatic, py::keep_alive<0, 1>())
-      .def("__mul__", py::overload_cast<int64_t>(&LinearExpr::Mul), arg("cst"),
+      .def("__sub__", &LinearExpr::SubDouble, arg("cst"),
            py::return_value_policy::automatic, py::keep_alive<0, 1>())
-      .def("__rmul__", py::overload_cast<double>(&LinearExpr::Mul), arg("cst"),
+      .def("__rsub__", &LinearExpr::RSubInt, arg("cst"),
            py::return_value_policy::automatic, py::keep_alive<0, 1>())
-      .def("__rmul__", py::overload_cast<int64_t>(&LinearExpr::Mul), arg("cst"),
+      .def("__rsub__", &LinearExpr::RSubDouble, arg("cst"),
            py::return_value_policy::automatic, py::keep_alive<0, 1>())
+      .def("__mul__", py::overload_cast<int64_t>(&LinearExpr::MulInt),
+           arg("cst"), py::return_value_policy::automatic,
+           py::keep_alive<0, 1>())
+      .def("__mul__", py::overload_cast<double>(&LinearExpr::MulDouble),
+           arg("cst"), py::return_value_policy::automatic,
+           py::keep_alive<0, 1>())
+      .def("__rmul__", py::overload_cast<int64_t>(&LinearExpr::MulInt),
+           arg("cst"), py::return_value_policy::automatic,
+           py::keep_alive<0, 1>())
+      .def("__rmul__", py::overload_cast<double>(&LinearExpr::MulDouble),
+           arg("cst"), py::return_value_policy::automatic,
+           py::keep_alive<0, 1>())
       .def("__neg__", &LinearExpr::Neg, py::return_value_policy::automatic,
            py::keep_alive<0, 1>())
       .def(
           "__eq__",
-          [](LinearExpr* expr, ExprOrValue rhs) {
-            if (rhs.int_value == std::numeric_limits<int64_t>::max() ||
-                rhs.int_value == std::numeric_limits<int64_t>::min()) {
-              throw_error(PyExc_ValueError,
-                          "== INT_MIN or INT_MAX is not supported");
+          [](LinearExpr* expr, LinearExpr* rhs) {
+            if (rhs == nullptr) {
+              throw_error(PyExc_TypeError,
+                          "Linear constraints do not accept None as argument.");
             }
             BoundedLinearExpression* result = expr->Eq(rhs);
             if (result == nullptr) {
               throw_error(PyExc_TypeError,
-                          "Linear constraint only accepts integer values "
+                          "Linear constraints only accept integer values "
+                          "and coefficients.");
+            }
+            return result;
+          },
+          py::return_value_policy::automatic, py::keep_alive<0, 1>(),
+          py::keep_alive<0, 2>())
+      .def(
+          "__eq__",
+          [](LinearExpr* expr, int64_t rhs) {
+            if (rhs == std::numeric_limits<int64_t>::max() ||
+                rhs == std::numeric_limits<int64_t>::min()) {
+              throw_error(PyExc_ValueError,
+                          "== INT_MIN or INT_MAX is not supported");
+            }
+            BoundedLinearExpression* result = expr->EqCst(rhs);
+            if (result == nullptr) {
+              throw_error(PyExc_TypeError,
+                          "Linear constraints only accept integer values "
+                          "and coefficients.");
+            }
+            return result;
+          },
+          py::return_value_policy::automatic, py::keep_alive<0, 1>())
+      .def(
+          "__ne__",
+          [](LinearExpr* expr, LinearExpr* rhs) {
+            if (rhs == nullptr) {
+              throw_error(PyExc_TypeError,
+                          "Linear constraints do not accept None as argument.");
+            }
+            BoundedLinearExpression* result = expr->Ne(rhs);
+            if (result == nullptr) {
+              throw_error(PyExc_TypeError,
+                          "Linear constraints only accept integer values "
                           "and coefficients.");
             }
             return result;
@@ -452,11 +494,27 @@ PYBIND11_MODULE(swig_helper, m) {
           py::keep_alive<0, 2>())
       .def(
           "__ne__",
-          [](LinearExpr* expr, ExprOrValue rhs) {
-            BoundedLinearExpression* result = expr->Ne(rhs);
+          [](LinearExpr* expr, int64_t rhs) {
+            BoundedLinearExpression* result = expr->NeCst(rhs);
             if (result == nullptr) {
               throw_error(PyExc_TypeError,
-                          "Linear constraint only accepts integer values "
+                          "Linear constraints only accept integer values "
+                          "and coefficients.");
+            }
+            return result;
+          },
+          py::return_value_policy::automatic, py::keep_alive<0, 1>())
+      .def(
+          "__le__",
+          [](LinearExpr* expr, LinearExpr* rhs) {
+            if (rhs == nullptr) {
+              throw_error(PyExc_TypeError,
+                          "Linear constraints do not accept None as argument.");
+            }
+            BoundedLinearExpression* result = expr->Le(rhs);
+            if (result == nullptr) {
+              throw_error(PyExc_TypeError,
+                          "Linear constraints only accept integer values "
                           "and coefficients.");
             }
             return result;
@@ -465,14 +523,30 @@ PYBIND11_MODULE(swig_helper, m) {
           py::keep_alive<0, 2>())
       .def(
           "__le__",
-          [](LinearExpr* expr, ExprOrValue rhs) {
-            if (rhs.int_value == std::numeric_limits<int64_t>::min()) {
+          [](LinearExpr* expr, int64_t rhs) {
+            if (rhs == std::numeric_limits<int64_t>::min()) {
               throw_error(PyExc_ArithmeticError, "<= INT_MIN is not supported");
             }
-            BoundedLinearExpression* result = expr->Le(rhs);
+            BoundedLinearExpression* result = expr->LeCst(rhs);
             if (result == nullptr) {
               throw_error(PyExc_TypeError,
-                          "Linear constraint only accepts integer values "
+                          "Linear constraints only accept integer values "
+                          "and coefficients.");
+            }
+            return result;
+          },
+          py::return_value_policy::automatic, py::keep_alive<0, 1>())
+      .def(
+          "__lt__",
+          [](LinearExpr* expr, LinearExpr* rhs) {
+            if (rhs == nullptr) {
+              throw_error(PyExc_TypeError,
+                          "Linear constraints do not accept None as argument.");
+            }
+            BoundedLinearExpression* result = expr->Lt(rhs);
+            if (result == nullptr) {
+              throw_error(PyExc_TypeError,
+                          "Linear constraints only accept integer values "
                           "and coefficients.");
             }
             return result;
@@ -481,14 +555,30 @@ PYBIND11_MODULE(swig_helper, m) {
           py::keep_alive<0, 2>())
       .def(
           "__lt__",
-          [](LinearExpr* expr, ExprOrValue rhs) {
-            if (rhs.int_value == std::numeric_limits<int64_t>::min()) {
+          [](LinearExpr* expr, int64_t rhs) {
+            if (rhs == std::numeric_limits<int64_t>::min()) {
               throw_error(PyExc_ArithmeticError, "< INT_MIN is not supported");
             }
-            BoundedLinearExpression* result = expr->Lt(rhs);
+            BoundedLinearExpression* result = expr->LtCst(rhs);
             if (result == nullptr) {
               throw_error(PyExc_TypeError,
-                          "Linear constraint only accepts integer values "
+                          "Linear constraints only accept integer values "
+                          "and coefficients.");
+            }
+            return result;
+          },
+          py::return_value_policy::automatic, py::keep_alive<0, 1>())
+      .def(
+          "__ge__",
+          [](LinearExpr* expr, LinearExpr* rhs) {
+            if (rhs == nullptr) {
+              throw_error(PyExc_TypeError,
+                          "Linear constraints do not accept None as argument.");
+            }
+            BoundedLinearExpression* result = expr->Ge(rhs);
+            if (result == nullptr) {
+              throw_error(PyExc_TypeError,
+                          "Linear constraints only accept integer values "
                           "and coefficients.");
             }
             return result;
@@ -497,14 +587,30 @@ PYBIND11_MODULE(swig_helper, m) {
           py::keep_alive<0, 2>())
       .def(
           "__ge__",
-          [](LinearExpr* expr, ExprOrValue rhs) {
-            if (rhs.int_value == std::numeric_limits<int64_t>::max()) {
+          [](LinearExpr* expr, int64_t rhs) {
+            if (rhs == std::numeric_limits<int64_t>::max()) {
               throw_error(PyExc_ArithmeticError, ">= INT_MAX is not supported");
             }
-            BoundedLinearExpression* result = expr->Ge(rhs);
+            BoundedLinearExpression* result = expr->GeCst(rhs);
             if (result == nullptr) {
               throw_error(PyExc_TypeError,
-                          "Linear constraint only accepts integer values "
+                          "Linear constraints only accept integer values "
+                          "and coefficients.");
+            }
+            return result;
+          },
+          py::return_value_policy::automatic, py::keep_alive<0, 1>())
+      .def(
+          "__gt__",
+          [](LinearExpr* expr, LinearExpr* rhs) {
+            if (rhs == nullptr) {
+              throw_error(PyExc_TypeError,
+                          "Linear constraints do not accept None as argument.");
+            }
+            BoundedLinearExpression* result = expr->Gt(rhs);
+            if (result == nullptr) {
+              throw_error(PyExc_TypeError,
+                          "Linear constraints only accept integer values "
                           "and coefficients.");
             }
             return result;
@@ -513,20 +619,19 @@ PYBIND11_MODULE(swig_helper, m) {
           py::keep_alive<0, 2>())
       .def(
           "__gt__",
-          [](LinearExpr* expr, ExprOrValue rhs) {
-            if (rhs.int_value == std::numeric_limits<int64_t>::max()) {
+          [](LinearExpr* expr, int64_t rhs) {
+            if (rhs == std::numeric_limits<int64_t>::max()) {
               throw_error(PyExc_ArithmeticError, "> INT_MAX is not supported");
             }
-            BoundedLinearExpression* result = expr->Gt(rhs);
+            BoundedLinearExpression* result = expr->GtCst(rhs);
             if (result == nullptr) {
               throw_error(PyExc_TypeError,
-                          "Linear constraint only accepts integer values "
+                          "Linear constraints only accept integer values "
                           "and coefficients.");
             }
             return result;
           },
-          py::return_value_policy::automatic, py::keep_alive<0, 1>(),
-          py::keep_alive<0, 2>())
+          py::return_value_policy::automatic, py::keep_alive<0, 1>())
       .def("__div__",
            [](LinearExpr* /*self*/, LinearExpr* /*other*/) {
              throw_error(PyExc_NotImplementedError,