diff --git a/vllm/entrypoints/openai/serving_responses.py b/vllm/entrypoints/openai/serving_responses.py
index 2ee8de5fb..dacf61dee 100644
--- a/vllm/entrypoints/openai/serving_responses.py
+++ b/vllm/entrypoints/openai/serving_responses.py
@@ -1000,11 +1000,6 @@ class OpenAIServingResponses(OpenAIServing):
                 # to add the tool call request to prev_outputs so that the
                 # parse_response_input can find the tool call request when
                 # parsing the tool call output.
-                if (
-                    isinstance(response_msg, dict)
-                    and response_msg.get("type") == "function_call"
-                ):
-                    response_msg = ResponseFunctionToolCall.model_validate(response_msg)
                 if isinstance(response_msg, ResponseFunctionToolCall):
                     prev_outputs.append(response_msg)
         return messages