refactor: keep the pre-existing cost-estimate comment and usage cost read out of the cleanup

Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>

litellm/proxy/management_endpoints/cost_tracking_settings.py

@@ -634,6 +634,7 @@ async def estimate_cost(
     # Pinning one moment keeps an off-peak window that opens mid-quote from pricing the totals on
     # one side of it and the reported rates on the other.
     with pinned_billing_time(current_billing_time()):
+        # Use completion_cost which handles all the logic including margins/discounts
         try:
             cost_per_request: Final = completion_cost(
                 completion_response=mock_response,

litellm/responses/streaming_iterator.py

@@ -1323,7 +1323,7 @@ def _stamp_responses_usage_cost(
     if usage_obj is None:
         return
     response_obj.usage = usage_obj  # rebind-ok: the stamped cost has to ride on the response the client receives
-    if isinstance(usage_obj.cost, (int, float)):
+    if isinstance(getattr(usage_obj, "cost", None), (int, float)):
         return
     try:
         cost: Final[float | None] = logging_obj._response_cost_calculator(result=response_obj)