diff --git a/mlx/backend/metal/kernels/reduction/reduce_row.h b/mlx/backend/metal/kernels/reduction/reduce_row.h
index 936d75bb52..b55c83f315 100644
--- a/mlx/backend/metal/kernels/reduction/reduce_row.h
+++ b/mlx/backend/metal/kernels/reduction/reduce_row.h
@@ -337,7 +337,8 @@ template <
 
   // lid.x * N_READS breaks the per_thread_row_reduce interface a bit. Maybe it
   // needs a small refactor.
-  in += elem_to_loc<IdxT>(out_idx, shape, strides, ndim) + lid.x * N_READS;
+  in +=
+      elem_to_loc<IdxT>(out_idx, shape, strides, ndim) + IdxT(lid.x) * N_READS;
 
   LoopedElemToLoc<NDIMS, IdxT, (NDIMS > 2)> loop(reduce_ndim);
   const device T* row;
diff --git a/python/tests/test_reduce.py b/python/tests/test_reduce.py
index 6ac8fc1504..164e2dd803 100644
--- a/python/tests/test_reduce.py
+++ b/python/tests/test_reduce.py
@@ -1,6 +1,5 @@
 # Copyright © 2023 Apple Inc.
 
-import unittest
 from itertools import combinations, permutations
 
 import mlx.core as mx
@@ -47,6 +46,16 @@ def test_expand_sums(self):
                                     np.allclose(z_npy, np.array(z_mlx), atol=1e-4)
                                 )
 
+    def test_row_reduce_negative_stride(self):
+        x_npy = np.arange(1, 131).reshape(2, 65)[::-1]
+        x_mlx = mx.arange(1, 131).reshape(2, 65)[::-1]
+
+        for op in ["sum", "max", "min", "mean", "var"]:
+            with self.subTest(op=op):
+                expected = getattr(np, op)(x_npy, axis=-1)
+                actual = getattr(mx, op)(x_mlx, axis=-1)
+                self.assertTrue(np.allclose(expected, actual))
+
     def test_dtypes(self):
         int_dtypes = [
             "int8",