diff --git a/python/tests/kernel/test_assignments.py b/python/tests/kernel/test_assignments.py
index 1551a5a6650..0650c965581 100644
--- a/python/tests/kernel/test_assignments.py
+++ b/python/tests/kernel/test_assignments.py
@@ -844,26 +844,27 @@ class MyTuple:
         l1: list[int]
         l2: list[int]
 
-    # FIXME: This test is flaky on py3.11 + arm64 + cuda12.6.
+    # NOTE: This test is flaky on py3.11 + arm64 + cuda12.6.
     # The struct argument marshaling is now correct (see PR #3879), but the
     # kernel-internal deep copy and list mutation still produce wrong results
     # intermittently on that configuration. Disabling until root cause is
     # identified.
-    # @cudaq.kernel
-    # def test11(t: MyTuple, size: int) -> list[int]:
-    #     l = [t.copy(deep=True) for _ in range(size)]
-    #     l[0].l1 = [2]
-    #     l[1].l2[0] = 3
-    #     res = [0 for _ in range(4 * len(l))]
-    #     for idx, item in enumerate(l):
-    #         res[4 * idx] = len(item.l1)
-    #         res[4 * idx + 1] = item.l1[0]
-    #         res[4 * idx + 2] = len(item.l2)
-    #         res[4 * idx + 3] = item.l2[0]
-    #     return res
-    #
-    # result = test11(MyTuple([1], [1]), 2)
-    # assert (result == [1, 2, 1, 1, 1, 1, 1, 3])
+    # See https://github.com/NVIDIA/cuda-quantum/issues/3923
+    @cudaq.kernel
+    def test11(t: MyTuple, size: int) -> list[int]:
+        l = [t.copy(deep=True) for _ in range(size)]
+        l[0].l1 = [2]
+        l[1].l2[0] = 3
+        res = [0 for _ in range(4 * len(l))]
+        for idx, item in enumerate(l):
+            res[4 * idx] = len(item.l1)
+            res[4 * idx + 1] = item.l1[0]
+            res[4 * idx + 2] = len(item.l2)
+            res[4 * idx + 3] = item.l2[0]
+        return res
+
+    result = test11(MyTuple([1], [1]), 2)
+    assert (result == [1, 2, 1, 1, 1, 1, 1, 3])
 
     with pytest.raises(RuntimeError) as e: