fix(responses): generate Responses-compatible IDs in Chat Completions bridge

The Chat Completions -> Responses API bridge reused chatcmpl-* IDs for
response and message output items. When bridged output from a non-OpenAI
provider (e.g. Claude via LiteLLM) is later sent as input to an OpenAI
Responses model, OpenAI rejects the request because message item IDs
must start with msg_.

Generate proper Responses-compatible IDs:
- resp_<uuid> for the top-level ResponsesAPIResponse
- msg_<uuid> for message output items
- img_<uuid> for image generation output items

Fixes #27333

litellm/responses/litellm_completion_transformation/transformation.py

@@ -2,6 +2,7 @@
 Handles transforming from Responses API -> LiteLLM completion  (Chat Completion API)
 """
 
+import uuid
 from collections.abc import Sequence
 from typing import Any, Dict, List, Literal, Optional, Set, Tuple, Union, cast
 
@@ -1657,7 +1658,7 @@ class LiteLLMCompletionResponsesConfig:
             finish_reason = choices[0].finish_reason
 
         responses_api_response: ResponsesAPIResponse = ResponsesAPIResponse(
-            id=chat_completion_response.id,
+            id=f"resp_{uuid.uuid4()}",
             created_at=chat_completion_response.created,
             model=chat_completion_response.model,
             object="response",
@@ -1874,7 +1875,7 @@ class LiteLLMCompletionResponsesConfig:
                 image_generation_items.append(
                     OutputImageGenerationCall(
                         type="image_generation_call",
-                        id=f"{chat_completion_response.id}_img_{idx}",
+                        id=f"img_{uuid.uuid4()}",
                         status=LiteLLMCompletionResponsesConfig._map_finish_reason_to_image_generation_status(
                             choice.finish_reason
                         ),
@@ -1950,7 +1951,7 @@ class LiteLLMCompletionResponsesConfig:
                 message_output_items.append(
                     GenericResponseOutputItem(
                         type="message",
-                        id=chat_completion_response.id,
+                        id=f"msg_{uuid.uuid4()}",
                         status=LiteLLMCompletionResponsesConfig._map_chat_completion_finish_reason_to_responses_status(
                             choice.finish_reason
                         ),

tests/test_litellm/responses/litellm_completion_transformation/test_image_generation_output.py

@@ -89,8 +89,8 @@ class TestExtractImageGenerationOutputItems:
         assert result[0].type == "image_generation_call"
         assert result[0].result == "IMG1"
         assert result[1].result == "IMG2"
-        assert result[0].id == "test_123_img_0"
-        assert result[1].id == "test_123_img_1"
+        assert result[0].id.startswith("img_")
+        assert result[1].id.startswith("img_")
         assert result[0].status == "completed"
 
     def test_returns_empty_for_no_images(self):

tests/test_litellm/responses/litellm_completion_transformation/test_response_id_prefixes.py

@@ -0,0 +1,115 @@
+"""
+Regression tests for Responses API bridge ID prefixes.
+
+The Chat Completions -> Responses bridge must generate Responses-compatible
+IDs (resp_*, msg_*) instead of reusing Chat Completions IDs (chatcmpl-*).
+Reusing chatcmpl-* IDs causes OpenAI to reject the request when bridged
+output is later sent back as Responses input.
+
+Regression test for https://github.com/BerriAI/litellm/issues/27333
+"""
+
+import os
+import sys
+
+sys.path.insert(0, os.path.abspath("../../.."))
+
+from litellm.responses.litellm_completion_transformation.transformation import (
+    LiteLLMCompletionResponsesConfig,
+)
+from litellm.types.utils import (
+    Choices,
+    Message,
+    ModelResponse,
+    Usage,
+)
+
+
+def _make_chat_completion_response(**overrides) -> ModelResponse:
+    defaults = dict(
+        id="chatcmpl-dfa2da3a-1586-4ff7-b64e-f59c692a5d11",
+        created=1717000000,
+        model="claude-3-5-sonnet-20241022",
+        object="chat.completion",
+        choices=[
+            Choices(
+                index=0,
+                finish_reason="stop",
+                message=Message(role="assistant", content="Hello"),
+            )
+        ],
+        usage=Usage(prompt_tokens=10, completion_tokens=5, total_tokens=15),
+    )
+    defaults.update(overrides)
+    return ModelResponse(**defaults)
+
+
+class TestResponseIdPrefixes:
+    """Bridged Responses output must use resp_*/msg_* ID prefixes."""
+
+    def test_response_id_uses_resp_prefix(self):
+        """Top-level response ID must start with resp_, not chatcmpl-."""
+        chat_response = _make_chat_completion_response()
+
+        result = LiteLLMCompletionResponsesConfig.transform_chat_completion_response_to_responses_api_response(
+            request_input="test",
+            responses_api_request={},
+            chat_completion_response=chat_response,
+        )
+
+        assert result.id.startswith(
+            "resp_"
+        ), f"Expected resp_* prefix, got: {result.id}"
+        assert not result.id.startswith("chatcmpl-")
+
+    def test_message_output_id_uses_msg_prefix(self):
+        """Message output item ID must start with msg_, not chatcmpl-."""
+        chat_response = _make_chat_completion_response()
+
+        result = LiteLLMCompletionResponsesConfig.transform_chat_completion_response_to_responses_api_response(
+            request_input="test",
+            responses_api_request={},
+            chat_completion_response=chat_response,
+        )
+
+        message_items = [
+            item for item in result.output if getattr(item, "type", None) == "message"
+        ]
+        assert len(message_items) > 0
+
+        for item in message_items:
+            assert item.id.startswith("msg_"), f"Expected msg_* prefix, got: {item.id}"
+            assert not item.id.startswith("chatcmpl-")
+
+    def test_response_and_message_ids_are_distinct(self):
+        """Response ID and message item ID must not be the same value."""
+        chat_response = _make_chat_completion_response()
+
+        result = LiteLLMCompletionResponsesConfig.transform_chat_completion_response_to_responses_api_response(
+            request_input="test",
+            responses_api_request={},
+            chat_completion_response=chat_response,
+        )
+
+        message_items = [
+            item for item in result.output if getattr(item, "type", None) == "message"
+        ]
+        for item in message_items:
+            assert result.id != item.id
+
+    def test_dict_input_also_gets_correct_prefixes(self):
+        """When chat_completion_response is passed as a dict, IDs still get correct prefixes."""
+        chat_response = _make_chat_completion_response()
+
+        result = LiteLLMCompletionResponsesConfig.transform_chat_completion_response_to_responses_api_response(
+            request_input="test",
+            responses_api_request={},
+            chat_completion_response=dict(chat_response),
+        )
+
+        assert result.id.startswith("resp_")
+        message_items = [
+            item for item in result.output if getattr(item, "type", None) == "message"
+        ]
+        for item in message_items:
+            assert item.id.startswith("msg_")