Merge pull request #14675 from tcx4c70/fix/response-price

fix: cost calculation for responses

Author: krrishdholakia

litellm/responses/utils.py

@@ -23,7 +23,7 @@
     ResponseText,
 )
 from litellm.types.responses.main import DecodedResponseId
-from litellm.types.utils import SpecialEnums, Usage
+from litellm.types.utils import PromptTokensDetails, SpecialEnums, Usage
 
 
 class ResponsesAPIRequestUtils:
@@ -375,8 +375,15 @@ def _transform_response_api_usage_to_chat_usage(
         )
         prompt_tokens: int = response_api_usage.input_tokens or 0
         completion_tokens: int = response_api_usage.output_tokens or 0
+        prompt_tokens_details: Optional[PromptTokensDetails] = None
+        if response_api_usage.input_tokens_details:
+            prompt_tokens_details = PromptTokensDetails(
+                cached_tokens=response_api_usage.input_tokens_details.cached_tokens,
+                audio_tokens=response_api_usage.input_tokens_details.audio_tokens,
+            )
         return Usage(
             prompt_tokens=prompt_tokens,
             completion_tokens=completion_tokens,
             total_tokens=prompt_tokens + completion_tokens,
-        )
+            prompt_tokens_details=prompt_tokens_details,
+        )

tests/test_litellm/responses/test_responses_utils.py

@@ -168,6 +168,7 @@ def test_transform_response_api_usage_to_chat_usage(self):
             "input_tokens": 10,
             "output_tokens": 20,
             "total_tokens": 30,
+            "input_tokens_details": {"cached_tokens": 2},
             "output_tokens_details": {"reasoning_tokens": 5},
         }
 
@@ -181,6 +182,7 @@ def test_transform_response_api_usage_to_chat_usage(self):
         assert result.prompt_tokens == 10
         assert result.completion_tokens == 20
         assert result.total_tokens == 30
+        assert result.prompt_tokens_details and result.prompt_tokens_details.cached_tokens == 2
 
     def test_transform_response_api_usage_with_none_values(self):
         """Test transformation handles None values properly"""