diff --git a/aeon/testing/estimator_checking/_yield_classification_checks.py b/aeon/testing/estimator_checking/_yield_classification_checks.py
index 0591119f7a..399747c2db 100644
--- a/aeon/testing/estimator_checking/_yield_classification_checks.py
+++ b/aeon/testing/estimator_checking/_yield_classification_checks.py
@@ -374,3 +374,10 @@ def check_classifier_output(estimator, datatype):
     # check predict proba (all classifiers have predict_proba by default)
     y_proba = estimator.predict_proba(FULL_TEST_DATA_DICT[datatype]["test"][0])
     _assert_predict_probabilities(y_proba, datatype, n_classes=len(unique_labels))
+
+    y_pred_proba_indices = np.argmax(y_proba, axis=1)
+    y_pred_proba = estimator.classes_[y_pred_proba_indices]
+
+    np.testing.assert_array_equal(
+        y_pred, y_pred_proba, err_msg="predict and predict_proba are not consistent"
+    )
diff --git a/aeon/testing/estimator_checking/_yield_clustering_checks.py b/aeon/testing/estimator_checking/_yield_clustering_checks.py
index 4e3940c489..6ed44212a8 100644
--- a/aeon/testing/estimator_checking/_yield_clustering_checks.py
+++ b/aeon/testing/estimator_checking/_yield_clustering_checks.py
@@ -141,6 +141,15 @@ def check_clusterer_output(estimator, datatype):
     assert isinstance(y_proba, np.ndarray)
     np.testing.assert_almost_equal(y_proba.sum(axis=1), 1, decimal=4)
 
+    # check predict and predict_proba have consistent outputs
+    y_pred_proba = np.argmax(y_proba, axis=1)
+
+    np.testing.assert_array_equal(
+        y_pred,
+        y_pred_proba,
+        err_msg="predict and predict_proba outputs are inconsistent",
+    )
+
 
 def check_clusterer_saving_loading_deep_learning(estimator_class, datatype):
     """Test Deep Clusterer saving."""