diff --git a/tests/python-gpu/test_gpu_updaters.py b/tests/python-gpu/test_gpu_updaters.py
index 48ac2f5e5..c1a076144 100644
--- a/tests/python-gpu/test_gpu_updaters.py
+++ b/tests/python-gpu/test_gpu_updaters.py
@@ -86,6 +86,7 @@ class TestGPUUpdaters:
     @settings(deadline=None)
     @pytest.mark.skipif(**tm.no_pandas())
     def test_categorical(self, rows, cols, rounds, cats):
+        pytest.xfail(reason='TestGPUUpdaters::test_categorical is flaky')
         self.run_categorical_basic(rows, cols, rounds, cats)
 
     def test_categorical_32_cat(self):
diff --git a/tests/python/test_with_dask.py b/tests/python/test_with_dask.py
index fe4a09556..f759c63bd 100644
--- a/tests/python/test_with_dask.py
+++ b/tests/python/test_with_dask.py
@@ -142,6 +142,9 @@ def test_dask_predict_shape_infer():
 
 @pytest.mark.parametrize("tree_method", ["hist", "approx"])
 def test_boost_from_prediction(tree_method):
+    if tree_method == 'approx':
+        pytest.xfail(reason='test_boost_from_prediction[approx] is flaky')
+
     from sklearn.datasets import load_breast_cancer
     X, y = load_breast_cancer(return_X_y=True)