diff --git a/litellm/cost_calculator.py b/litellm/cost_calculator.py
index 28cda2daa77..42a7fb0bb02 100644
--- a/litellm/cost_calculator.py
+++ b/litellm/cost_calculator.py
@@ -1505,6 +1505,12 @@ def _response_time_ms_for_cost(
     return 0.0
 
 
+def _breakdown_total(litellm_logging_obj: LitellmLoggingObject | None) -> float | None:
+    breakdown: Final[object] = getattr(litellm_logging_obj, "cost_breakdown", None)
+    total: Final[object] = breakdown.get("total_cost") if isinstance(breakdown, Mapping) else None
+    return float(total) if isinstance(total, (int, float)) and not isinstance(total, bool) else None
+
+
 def _adjusted_provider_billed_cost(
     reported_cost: float,
     custom_llm_provider: str | None,
@@ -2334,14 +2340,15 @@ def response_cost_calculator(
                     )
                     provider_response_cost: Final = get_response_cost_from_hidden_params(hidden_params)
                     if provider_response_cost is not None and not custom_pricing:
-                        _store_cost_breakdown_in_logging_obj(
-                            litellm_logging_obj=litellm_logging_obj,
-                            prompt_tokens_cost_usd_dollar=0.0,
-                            completion_tokens_cost_usd_dollar=provider_response_cost,
-                            cost_for_built_in_tools_cost_usd_dollar=0.0,
-                            total_cost_usd_dollar=provider_response_cost,
-                            original_cost=provider_response_cost,
-                        )
+                        if _breakdown_total(litellm_logging_obj) != provider_response_cost:
+                            _store_cost_breakdown_in_logging_obj(
+                                litellm_logging_obj=litellm_logging_obj,
+                                prompt_tokens_cost_usd_dollar=0.0,
+                                completion_tokens_cost_usd_dollar=provider_response_cost,
+                                cost_for_built_in_tools_cost_usd_dollar=0.0,
+                                total_cost_usd_dollar=provider_response_cost,
+                                original_cost=provider_response_cost,
+                            )
                         return provider_response_cost
 
             response_cost = completion_cost(
diff --git a/litellm/litellm_core_utils/core_helpers.py b/litellm/litellm_core_utils/core_helpers.py
index 83e9de1378b..31230d4e8d1 100644
--- a/litellm/litellm_core_utils/core_helpers.py
+++ b/litellm/litellm_core_utils/core_helpers.py
@@ -1028,7 +1028,8 @@ def mark_usage_cost_as_litellm_bill(response: object, bill: float) -> None:
     so ``billed_usage_cost`` does not read it back as the upstream's charge on a later pricing pass."""
     hidden: Final[object] = getattr(response, "_hidden_params", None)
     if isinstance(hidden, dict):
-        cast(dict[str, object], hidden).update({"response_cost": bill, LITELLM_BILLED_USAGE_COST: bill})
+        hidden_params: Final = cast(dict[str, object], hidden)  # cast-ok: hidden params ride on the response as a dict
+        hidden_params.update({"response_cost": bill, LITELLM_BILLED_USAGE_COST: bill})
 
 
 def _usage_cost_is_litellm_bill(response: object) -> bool: