From 5e8dd8d5e07d392a21ded991627ca07989f37ee5 Mon Sep 17 00:00:00 2001
From: shrey kharbanda <shreshth@berri.ai>
Date: Mon, 28 Sep 2026 20:25:31 +0000
Subject: [PATCH] fix(logging): keep the uncached price probe's restores to the
 state it can leak

---
 litellm/litellm_core_utils/litellm_logging.py |  3 --
 .../test_litellm_logging.py                   | 49 +++++++++++++------
 2 files changed, 35 insertions(+), 17 deletions(-)

diff --git a/litellm/litellm_core_utils/litellm_logging.py b/litellm/litellm_core_utils/litellm_logging.py
index ca563da9cca..ab7b0965d92 100644
--- a/litellm/litellm_core_utils/litellm_logging.py
+++ b/litellm/litellm_core_utils/litellm_logging.py
@@ -1976,10 +1976,8 @@ class Logging(LiteLLMLoggingBaseClass):
             list,
         ],
     ) -> float | None:
-        """Price a response as if it had missed the cache, without leaving the probe's side effects on this call"""
         previous_cost_breakdown: Final = self.cost_breakdown
         previous_billed_token_rates: Final = self.billed_token_rates
-        previous_zero_cost_warned: Final = self.zero_cost_warned
         debug_key: Final = "response_cost_failure_debug_information"
         debug_missing: Final = object()
         debug_before: Final = self.model_call_details.get(debug_key, debug_missing)
@@ -1988,7 +1986,6 @@ class Logging(LiteLLMLoggingBaseClass):
         finally:
             self.cost_breakdown = previous_cost_breakdown
             self.billed_token_rates = previous_billed_token_rates
-            self.zero_cost_warned = previous_zero_cost_warned
             if debug_before is debug_missing:
                 self.model_call_details.pop(debug_key, None)
             else:
diff --git a/tests/unit/litellm_core_utils/test_litellm_logging.py b/tests/unit/litellm_core_utils/test_litellm_logging.py
index 102a0fdd445..5334161fa10 100644
--- a/tests/unit/litellm_core_utils/test_litellm_logging.py
+++ b/tests/unit/litellm_core_utils/test_litellm_logging.py
@@ -9092,7 +9092,7 @@ def test_uncached_response_cost_recomputes_price_past_a_zero_stamp():
 
 
 def test_uncached_response_cost_leaves_no_pricing_state_on_the_call():
-    """The uncached price probe restores every field the priced path writes on the logging object."""
+    """The uncached price probe leaves no cost breakdown or billed rates behind on the cache-hit call."""
     deployment_id: Final = "lit8679-probe-state-deployment"
     litellm.register_model(
         model_cost={
@@ -9113,23 +9113,12 @@ def test_uncached_response_cost_leaves_no_pricing_state_on_the_call():
             usage=litellm.Usage(prompt_tokens=10, completion_tokens=20, total_tokens=30),
         )
         cached_response._hidden_params = {"cache_hit": True, "cache_key": "lit8679-key", "response_cost": 0.0}
-        state_before: Final = (
-            logging_obj.cost_breakdown,
-            logging_obj.billed_token_rates,
-            logging_obj.zero_cost_warned,
-            "response_cost_failure_debug_information" in logging_obj.model_call_details,
-        )
+        assert (logging_obj.cost_breakdown, logging_obj.billed_token_rates) == (None, None)
 
         cost: Final = logging_obj._uncached_response_cost(result=cached_response)
 
         assert cost is not None and cost > 0.0
-        assert (
-            logging_obj.cost_breakdown,
-            logging_obj.billed_token_rates,
-            logging_obj.zero_cost_warned,
-            "response_cost_failure_debug_information" in logging_obj.model_call_details,
-        ) == state_before
-        assert state_before == (None, None, False, False)
+        assert (logging_obj.cost_breakdown, logging_obj.billed_token_rates) == (None, None)
     finally:
         litellm.model_cost.pop(deployment_id, None)
 
@@ -9186,3 +9175,35 @@ def test_payload_saved_cache_cost_keeps_uncached_price_when_result_stamped_zero(
         assert logging_obj.cost_breakdown is None
     finally:
         litellm.model_cost.pop(deployment_id, None)
+
+
+def test_uncached_response_cost_leaves_no_failure_debug_info_for_unpriced_model():
+    """A probe that cannot price the model returns None without stamping the pricing failure on the cache-hit call."""
+    logging_obj: Final = LitellmLogging(
+        model="lit8679-model-with-no-price",
+        messages=[{"role": "user", "content": "Hi"}],
+        stream=False,
+        call_type="completion",
+        start_time=time.time(),
+        litellm_call_id="lit8679-unpriced-probe",
+        function_id="fn",
+    )
+    logging_obj.update_environment_variables(
+        model="lit8679-model-with-no-price",
+        user="",
+        optional_params={},
+        litellm_params={"metadata": {}},
+        custom_llm_provider="openai",
+    )
+    cached_response: Final = ModelResponse(
+        model="lit8679-model-with-no-price",
+        choices=[litellm.Choices(message=litellm.Message(role="assistant", content="cached"))],
+        usage=litellm.Usage(prompt_tokens=10, completion_tokens=20, total_tokens=30),
+    )
+    cached_response._hidden_params = {"cache_hit": True, "cache_key": "lit8679-key", "response_cost": 0.0}
+    assert "response_cost_failure_debug_information" not in logging_obj.model_call_details
+
+    cost: Final = logging_obj._uncached_response_cost(result=cached_response)
+
+    assert cost is None
+    assert "response_cost_failure_debug_information" not in logging_obj.model_call_details