diff --git a/litellm/proxy/common_request_processing.py b/litellm/proxy/common_request_processing.py
index 4349ec130c4..cb93c98cfae 100644
--- a/litellm/proxy/common_request_processing.py
+++ b/litellm/proxy/common_request_processing.py
@@ -1427,9 +1427,14 @@ class ProxyBaseLLMRequestProcessing:
 
         response = responses[1]
 
-        # GH#30566: overhead for non-chat-completions routes
+        # GH#30566: overhead for non-chat-completions routes.
+        # Skip chat routes (acompletion/completion) because the SDK already
+        # sets litellm_overhead_time_ms in litellm.utils.completion().
         _hidden_params = getattr(response, "_hidden_params", {}) or {}
-        if not _hidden_params.get("litellm_overhead_time_ms"):
+        if (
+            not _hidden_params.get("litellm_overhead_time_ms")
+            and route_type not in ("acompletion", "completion")
+        ):
             end_time = datetime.now()
             _logging_obj = self.data.get("litellm_logging_obj")
             if _logging_obj is not None: