diff --git a/llama_stack/providers/inline/agents/meta_reference/responses/utils.py b/llama_stack/providers/inline/agents/meta_reference/responses/utils.py
index 847cf0265..7aaeb4cd5 100644
--- a/llama_stack/providers/inline/agents/meta_reference/responses/utils.py
+++ b/llama_stack/providers/inline/agents/meta_reference/responses/utils.py
@@ -157,8 +157,10 @@ async def convert_response_input_to_chat_messages(
                         f"Llama Stack OpenAI Responses does not yet support message role '{input_item.role}' in this context"
                     )
                 messages.append(message_type(content=content))
-        for result in tool_call_results.values():
-            messages.append(result)
+        if len(tool_call_results):
+            raise ValueError(
+                f"Received function_call_output(s) with call_id(s) {tool_call_results.keys()}, but no corresponding function_call"
+            )
     else:
         messages.append(OpenAIUserMessageParam(content=input))
     return messages
diff --git a/tests/unit/providers/agents/meta_reference/test_response_conversion_utils.py b/tests/unit/providers/agents/meta_reference/test_response_conversion_utils.py
index 985a0e257..187540f82 100644
--- a/tests/unit/providers/agents/meta_reference/test_response_conversion_utils.py
+++ b/tests/unit/providers/agents/meta_reference/test_response_conversion_utils.py
@@ -115,18 +115,27 @@ class TestConvertResponseInputToChatMessages:
 
     async def test_convert_function_tool_call_output(self):
         input_items = [
+            OpenAIResponseOutputMessageFunctionToolCall(
+                call_id="call_123",
+                name="test_function",
+                arguments='{"param": "value"}',
+            ),
             OpenAIResponseInputFunctionToolCallOutput(
                 output="Tool output",
                 call_id="call_123",
-            )
+            ),
         ]
 
         result = await convert_response_input_to_chat_messages(input_items)
 
-        assert len(result) == 1
-        assert isinstance(result[0], OpenAIToolMessageParam)
-        assert result[0].content == "Tool output"
-        assert result[0].tool_call_id == "call_123"
+        assert len(result) == 2
+        assert isinstance(result[0], OpenAIAssistantMessageParam)
+        assert result[0].tool_calls[0].id == "call_123"
+        assert result[0].tool_calls[0].function.name == "test_function"
+        assert result[0].tool_calls[0].function.arguments == '{"param": "value"}'
+        assert isinstance(result[1], OpenAIToolMessageParam)
+        assert result[1].content == "Tool output"
+        assert result[1].tool_call_id == "call_123"
 
     async def test_convert_function_tool_call(self):
         input_items = [