diff --git a/doc/source/whatsnew/v1.5.0.rst b/doc/source/whatsnew/v1.5.0.rst
index 358d9447b131d..82694de531db6 100644
--- a/doc/source/whatsnew/v1.5.0.rst
+++ b/doc/source/whatsnew/v1.5.0.rst
@@ -500,6 +500,7 @@ Conversion
 - Bug in :func:`array` with ``FloatingDtype`` and values containing float-castable strings incorrectly raising (:issue:`45424`)
 - Bug when comparing string and datetime64ns objects causing ``OverflowError`` exception. (:issue:`45506`)
 - Bug in metaclass of generic abstract dtypes causing :meth:`DataFrame.apply` and :meth:`Series.apply` to raise for the built-in function ``type`` (:issue:`46684`)
+- Bug in :meth:`DataFrame.to_dict` for ``orient="list"`` or ``orient="index"`` was not returning native types (:issue:`46751`)
 
 Strings
 ^^^^^^^
diff --git a/pandas/core/frame.py b/pandas/core/frame.py
index ceede5fdb5577..061b01b6d169a 100644
--- a/pandas/core/frame.py
+++ b/pandas/core/frame.py
@@ -1913,7 +1913,9 @@ def to_dict(self, orient: str = "dict", into=dict):
             return into_c((k, v.to_dict(into)) for k, v in self.items())
 
         elif orient == "list":
-            return into_c((k, v.tolist()) for k, v in self.items())
+            return into_c(
+                (k, list(map(maybe_box_native, v.tolist()))) for k, v in self.items()
+            )
 
         elif orient == "split":
             return into_c(
@@ -1964,7 +1966,7 @@ def to_dict(self, orient: str = "dict", into=dict):
             if not self.index.is_unique:
                 raise ValueError("DataFrame index must be unique for orient='index'.")
             return into_c(
-                (t[0], dict(zip(self.columns, t[1:])))
+                (t[0], dict(zip(self.columns, map(maybe_box_native, t[1:]))))
                 for t in self.itertuples(name=None)
             )
 
diff --git a/pandas/tests/frame/methods/test_to_dict.py b/pandas/tests/frame/methods/test_to_dict.py
index 31ea3e582eeb2..6d5c32cae7368 100644
--- a/pandas/tests/frame/methods/test_to_dict.py
+++ b/pandas/tests/frame/methods/test_to_dict.py
@@ -344,3 +344,80 @@ def test_to_dict_orient_tight(self, index, columns):
         roundtrip = DataFrame.from_dict(df.to_dict(orient="tight"), orient="tight")
 
         tm.assert_frame_equal(df, roundtrip)
+
+    @pytest.mark.parametrize(
+        "orient",
+        ["dict", "list", "split", "records", "index", "tight"],
+    )
+    @pytest.mark.parametrize(
+        "data,expected_types",
+        (
+            (
+                {
+                    "a": [np.int64(1), 1, np.int64(3)],
+                    "b": [np.float64(1.0), 2.0, np.float64(3.0)],
+                    "c": [np.float64(1.0), 2, np.int64(3)],
+                    "d": [np.float64(1.0), "a", np.int64(3)],
+                    "e": [np.float64(1.0), ["a"], np.int64(3)],
+                    "f": [np.float64(1.0), ("a",), np.int64(3)],
+                },
+                {
+                    "a": [int, int, int],
+                    "b": [float, float, float],
+                    "c": [float, float, float],
+                    "d": [float, str, int],
+                    "e": [float, list, int],
+                    "f": [float, tuple, int],
+                },
+            ),
+            (
+                {
+                    "a": [1, 2, 3],
+                    "b": [1.1, 2.2, 3.3],
+                },
+                {
+                    "a": [int, int, int],
+                    "b": [float, float, float],
+                },
+            ),
+        ),
+    )
+    def test_to_dict_returns_native_types(self, orient, data, expected_types):
+        # GH 46751
+        # Tests we get back native types for all orient types
+        df = DataFrame(data)
+        result = df.to_dict(orient)
+        if orient == "dict":
+            assertion_iterator = (
+                (i, key, value)
+                for key, index_value_map in result.items()
+                for i, value in index_value_map.items()
+            )
+        elif orient == "list":
+            assertion_iterator = (
+                (i, key, value)
+                for key, values in result.items()
+                for i, value in enumerate(values)
+            )
+        elif orient in {"split", "tight"}:
+            assertion_iterator = (
+                (i, key, result["data"][i][j])
+                for i in result["index"]
+                for j, key in enumerate(result["columns"])
+            )
+        elif orient == "records":
+            assertion_iterator = (
+                (i, key, value)
+                for i, record in enumerate(result)
+                for key, value in record.items()
+            )
+        elif orient == "index":
+            assertion_iterator = (
+                (i, key, value)
+                for i, record in result.items()
+                for key, value in record.items()
+            )
+
+        for i, key, value in assertion_iterator:
+            assert value == data[key][i]
+            assert type(value) is expected_types[key][i]