diff --git a/src/challenger/input_parser.star b/src/challenger/input_parser.star
index 5f06e73b..935d9564 100644
--- a/src/challenger/input_parser.star
+++ b/src/challenger/input_parser.star
@@ -64,7 +64,7 @@ def _parse_instance(challenger_args, challenger_name, l2s_params):
 
     challenger_params["labels"] = {
         "op.kind": "challenger",
-        "op.network.id": ",".join([str(network_id) for network_id in network_ids]),
+        "op.network.id": "-".join([str(network_id) for network_id in network_ids]),
     }
 
     # Now we make sure to cover the prestate arg combinations
diff --git a/test/challenger/input_parser_test.star b/test/challenger/input_parser_test.star
index f0bd2166..db4345f9 100644
--- a/test/challenger/input_parser_test.star
+++ b/test/challenger/input_parser_test.star
@@ -39,7 +39,7 @@ def test_challenger_input_parser_default_args(plan):
         service_name="op-challenger-challenger-1000-2000",
         enabled=True,
         image="us-docker.pkg.dev/oplabs-tools-artifacts/images/op-challenger:develop",
-        labels={"op.kind": "challenger", "op.network.id": "1000,2000"},
+        labels={"op.kind": "challenger", "op.network.id": "1000-2000"},
         extra_params=[],
         participants=[1000, 2000],
         cannon_prestate_path="",