diff --git a/vllm/distributed/kv_transfer/kv_connector/v1/decode_bench_connector.py b/vllm/distributed/kv_transfer/kv_connector/v1/decode_bench_connector.py
index e1f4b86b81bc7..c6297eba6ae50 100644
--- a/vllm/distributed/kv_transfer/kv_connector/v1/decode_bench_connector.py
+++ b/vllm/distributed/kv_transfer/kv_connector/v1/decode_bench_connector.py
@@ -116,6 +116,15 @@ class DecodeBenchConnector(KVConnectorBase_V1):
         assert self.connector_scheduler is not None
         return self.connector_scheduler.build_connector_meta(scheduler_output)
 
+    def request_finished(
+        self,
+        request: "Request",
+        block_ids: list[int],
+    ) -> tuple[bool, Optional[dict[str, Any]]]:
+        assert self.connector_scheduler is not None
+        self.connector_scheduler.request_finished(request)
+        return False, None
+
 
 class DecodeBenchConnectorScheduler:
     """Scheduler-side implementation for DecodeBenchConnector."""
@@ -211,6 +220,12 @@ class DecodeBenchConnectorScheduler:
 
         return meta
 
+    def request_finished(self, request: "Request"):
+        """
+        Called when a request has finished. Clean up any state.
+        """
+        self._filled_requests.discard(request.request_id)
+
 
 class DecodeBenchConnectorWorker:
     """Worker-side implementation for DecodeBenchConnector."""