diff --git a/litellm/proxy/management_endpoints/cost_tracking_settings.py b/litellm/proxy/management_endpoints/cost_tracking_settings.py
index a679e79f8d4..dc0da63555f 100644
--- a/litellm/proxy/management_endpoints/cost_tracking_settings.py
+++ b/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,
diff --git a/litellm/responses/streaming_iterator.py b/litellm/responses/streaming_iterator.py
index a8dd8c2d8cd..40ff88fc557 100644
--- a/litellm/responses/streaming_iterator.py
+++ b/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)