diff --git a/qutip/core/superop_reps.py b/qutip/core/superop_reps.py
index 4413cfafb5..029aeab2cc 100644
--- a/qutip/core/superop_reps.py
+++ b/qutip/core/superop_reps.py
@@ -166,9 +166,10 @@ def kraus_to_choi(kraus_ops: list[Qobj]) -> Qobj:
     choi_dims = [kraus_ops[0].dims] * 2
     # transform a list of Qobj matrices list[sum_ij k_ij |i><j|]
     # into an array of array vectors sum_ij k_ij |i, j>> = sum_I k_I |I>>
-    kraus_vectors = np.asarray(
-        [np.reshape(kraus_op.full(), len_op, "F") for kraus_op in kraus_ops]
-    )
+    kraus_vectors = np.asarray([
+        np.reshape(kraus_op.full(), len_op, order="F")
+        for kraus_op in kraus_ops
+    ])
     # sum_{I} |k_I|^2 |I>><<I|
     choi_array = np.tensordot(
         kraus_vectors, kraus_vectors.conj(), axes=([0], [0])
@@ -205,7 +206,9 @@ def _super_tofrom_choi(q_oper):
     d1 = np.prod(flatten(new_dims[1]))
     s0 = np.prod(dims[0][0])
     s1 = np.prod(dims[1][1])
-    data = data.reshape([s0, s1, s0, s1]).transpose(3, 1, 2, 0).reshape(d0, d1)
+    data = (
+        data.reshape([s0, s1, s0, s1]).transpose(3, 1, 2, 0).reshape([d0, d1])
+    )
     return Qobj(data,
                 dims=new_dims,
                 superrep='super' if q_oper.superrep == 'choi' else 'choi',
diff --git a/qutip/partial_transpose.py b/qutip/partial_transpose.py
index 919bfa9e63..01fd5dfb68 100644
--- a/qutip/partial_transpose.py
+++ b/qutip/partial_transpose.py
@@ -55,7 +55,7 @@ def _partial_transpose_dense(rho, mask):
     Very fast for dense problems.
     """
     nsys = len(mask)
-    pt_dims = np.arange(2 * nsys).reshape(2, nsys).T
+    pt_dims = np.arange(2 * nsys).reshape([2, nsys]).T
     pt_idx = np.concatenate([[pt_dims[n, mask[n]] for n in range(nsys)],
                             [pt_dims[n, 1 - mask[n]] for n in range(nsys)]])
 
diff --git a/qutip/solver/sode/_noise.py b/qutip/solver/sode/_noise.py
index eec07a1485..19cac96825 100644
--- a/qutip/solver/sode/_noise.py
+++ b/qutip/solver/sode/_noise.py
@@ -65,7 +65,7 @@ def __init__(self, noise, tlist, n_sc_ops, heterodyne, is_measurement):
                     "Noise is not of the expected shape: "
                     f"{(n_sc_ops/2, 2, len(tlist)-1)}"
                 )
-            noise = np.reshape(noise, (n_sc_ops, len(tlist)-1), "C")
+            noise = np.reshape(noise, (n_sc_ops, len(tlist)-1), order="C")
         else:
             if noise.shape != (n_sc_ops, len(tlist)-1):
                 raise ValueError(
@@ -107,7 +107,7 @@ def dw(self, dt):
         Ito integral I(i).
         """
         N = int(np.round(dt / self.dt))
-        return self.noise.reshape(-1, N, self.num).sum(axis=1)
+        return self.noise.reshape([-1, N, self.num]).sum(axis=1)
 
     def dz(self, dt):
         """
@@ -117,7 +117,7 @@ def dz(self, dt):
         return (
             np.einsum(
                 "ijk,j->ik",
-                self.noise.reshape(-1, N, self.num),
+                self.noise.reshape([-1, N, self.num]),
                 np.arange(N - 0.5, 0, -1),
             )
             * self.dt
@@ -132,11 +132,11 @@ def dW(self, dt):
         if noise.shape[0] % N:
             noise = noise[: -(noise.shape[0] % N)]
         out = np.empty((noise.shape[0] // N, 2, self.num), dtype=float)
-        out[:, 0, :] = noise.reshape(-1, N, self.num).sum(axis=1)
+        out[:, 0, :] = noise.reshape([-1, N, self.num]).sum(axis=1)
         out[:, 1, :] = (
             np.einsum(
                 "ijk,j->ik",
-                self.noise.reshape(-1, N, self.num),
+                self.noise.reshape([-1, N, self.num]),
                 np.arange(N - 0.5, 0, -1),
             )
             * self.dt
diff --git a/qutip/tests/core/data/test_expect.py b/qutip/tests/core/data/test_expect.py
index c43da02def..47b8c79ce0 100644
--- a/qutip/tests/core/data/test_expect.py
+++ b/qutip/tests/core/data/test_expect.py
@@ -49,7 +49,7 @@ class TestExpectSuper(BinaryOpMixin):
     def op_numpy(self, op, state):
         n = np.sqrt(state.shape[0]).astype(int)
         out_shape = (n, n)
-        return np.trace(np.reshape(op@state, newshape=out_shape))
+        return np.trace(np.reshape(op@state, out_shape))
 
     _dim = 100
     _super_ket = pytest.param((_dim, 1), id="super_ket")
diff --git a/qutip/tests/core/data/test_reshape.py b/qutip/tests/core/data/test_reshape.py
index 5f23b06b83..f9e2947baa 100644
--- a/qutip/tests/core/data/test_reshape.py
+++ b/qutip/tests/core/data/test_reshape.py
@@ -20,7 +20,7 @@ def op_numpy(self, matrix):
 class TestColumnStack(UnaryOpMixin):
     def op_numpy(self, matrix):
         out_shape = (matrix.shape[0]*matrix.shape[1], 1)
-        return np.reshape(matrix, newshape=out_shape, order='F')
+        return np.reshape(matrix, out_shape, order='F')
 
     specialisations = [
         pytest.param(data.column_stack_csr, CSR, CSR),
@@ -32,7 +32,7 @@ def op_numpy(self, matrix):
 class TestColumnUnstack(UnaryOpMixin):
     def op_numpy(self, matrix, rows):
         out_shape = (rows, matrix.shape[0]*matrix.shape[1]//rows)
-        return np.reshape(matrix, newshape=out_shape, order='F')
+        return np.reshape(matrix, out_shape, order='F')
 
     shapes = [
         (pytest.param((10, 1), id="ket"), ),
@@ -83,7 +83,7 @@ def generate_incorrect_rows_raises(self, metafunc):
 class TestReshape(UnaryOpMixin):
     def op_numpy(self, matrix, rows, columns):
         out_shape = (rows, columns)
-        return np.reshape(matrix, newshape=out_shape, order='C')
+        return np.reshape(matrix, out_shape, order='C')
 
     # All matrices should have the same number of elements in total, so we can
     # use the same (rows, columns) parametrisation for each input.