diff --git a/sqlmesh/core/macros.py b/sqlmesh/core/macros.py
index ec5b2567f4..a72bf4605a 100644
--- a/sqlmesh/core/macros.py
+++ b/sqlmesh/core/macros.py
@@ -361,8 +361,37 @@ def evaluate(self, node: MacroFunc) -> exp.Expression | t.List[exp.Expression] |
             return None
 
         if isinstance(result, (tuple, list)):
-            return [self.parse_one(item) for item in result if item is not None]
-        return self.parse_one(result)
+            result = [self.parse_one(item) for item in result if item is not None]
+
+            if (
+                len(result) == 1
+                and isinstance(result[0], (exp.Array, exp.Tuple))
+                and node.find_ancestor(MacroFunc)
+            ):
+                """
+                if:
+                 - the output of evaluating this node is being passed as an argument to another macro function
+                 - and that output is something that _norm_var_arg_lambda() will unpack into varargs
+                   > (a list containing a single item of type exp.Tuple/exp.Array)
+                then we will get inconsistent behaviour depending on if this node emits a list with a single item vs multiple items.
+                
+                In the first case, emitting a list containing a single array item will cause that array to get unpacked and its *members* passed to the calling macro
+                In the second case, emitting a list containing multiple array items will cause each item to get passed as-is to the calling macro
+                
+                To prevent this inconsistency, we wrap this node output in an exp.Array so that _norm_var_arg_lambda() can "unpack" that into the
+                actual argument we want to pass to the parent macro function
+                
+                Note we only do this for evaluation results that get passed as an argument to another macro, because when the final
+                result is given to something like SELECT, we still want that to be unpacked into a list of items like:
+                 - SELECT ARRAY(1), ARRAY(2)
+                rather than a single item like:
+                 - SELECT ARRAY(ARRAY(1), ARRAY(2))                
+                """
+                result = [exp.Array(expressions=result)]
+        else:
+            result = self.parse_one(result)
+
+        return result
 
     def eval_expression(self, node: t.Any) -> t.Any:
         """Converts a SQLGlot expression into executable Python code and evals it.
diff --git a/tests/core/test_macros.py b/tests/core/test_macros.py
index 0e3615d6c0..77d8fb84ae 100644
--- a/tests/core/test_macros.py
+++ b/tests/core/test_macros.py
@@ -363,11 +363,24 @@ def test_ast_correctness(macro_evaluator):
             "SELECT column LIKE a OR column LIKE b OR column LIKE c",
             {},
         ),
+        ("SELECT @REDUCE([1], (x, y) -> x + y)", "SELECT 1", {}),
+        ("SELECT @REDUCE([1, 2], (x, y) -> x + y)", "SELECT 1 + 2", {}),
+        ("SELECT @REDUCE([[1]], (x, y) -> x + y)", "SELECT ARRAY(1)", {}),
+        ("SELECT @REDUCE([[1, 2]], (x, y) -> x + y)", "SELECT ARRAY(1, 2)", {}),
         (
             """select @EACH([a, b, c], x -> column like x AS @SQL('@{x}_y', 'Identifier')), @x""",
             "SELECT column LIKE a AS a_y, column LIKE b AS b_y, column LIKE c AS c_y, '3'",
             {"x": "3"},
         ),
+        ("SELECT @EACH([1], a -> [@a])", "SELECT ARRAY(1)", {}),
+        ("SELECT @EACH([1, 2], a -> [@a])", "SELECT ARRAY(1), ARRAY(2)", {}),
+        ("SELECT @REDUCE(@EACH([1], a -> [@a]), (x, y) -> x + y)", "SELECT ARRAY(1)", {}),
+        (
+            "SELECT @REDUCE(@EACH([1, 2], a -> [@a]), (x, y) -> x + y)",
+            "SELECT ARRAY(1) + ARRAY(2)",
+            {},
+        ),
+        ("SELECT @REDUCE([[1],[2]], (x, y) -> x + y)", "SELECT ARRAY(1) + ARRAY(2)", {}),
         (
             """@WITH(@do_with) all_cities as (select * from city) select all_cities""",
             "WITH all_cities AS (SELECT * FROM city) SELECT all_cities",