diff --git a/bigframes/session/__init__.py b/bigframes/session/__init__.py
index 0f7953d3d4..da6dd4fb20 100644
--- a/bigframes/session/__init__.py
+++ b/bigframes/session/__init__.py
@@ -1833,6 +1833,8 @@ def _start_query(
         Starts BigQuery query job and waits for results.
         """
         job_config = self._prepare_query_job_config(job_config)
+        if not self._strictly_ordered:
+            job_config.labels = {"bigframes-mode": "unordered"}
         try:
             return bigframes.session._io.bigquery.start_query_with_client(
                 self,
diff --git a/tests/system/small/test_unordered.py b/tests/system/small/test_unordered.py
index 36bf2a2585..f2ed6d2a01 100644
--- a/tests/system/small/test_unordered.py
+++ b/tests/system/small/test_unordered.py
@@ -20,6 +20,15 @@
 from tests.system.utils import assert_pandas_df_equal, skip_legacy_pandas
 
 
+def test_unordered_mode_job_label(unordered_session):
+    pd_df = pd.DataFrame({"a": [1, 2, 3], "b": [4, 5, 6]}, dtype=pd.Int64Dtype())
+    df = bpd.DataFrame(pd_df, session=unordered_session)
+    df.to_pandas()
+    job_labels = df.query_job.labels  # type:ignore
+    assert "bigframes-mode" in job_labels
+    assert job_labels["bigframes-mode"] == "unordered"
+
+
 def test_unordered_mode_cache_aggregate(unordered_session):
     pd_df = pd.DataFrame({"a": [1, 2, 3], "b": [4, 5, 6]}, dtype=pd.Int64Dtype())
     df = bpd.DataFrame(pd_df, session=unordered_session)