fix(bedrock): keep provider response headers on the runtime chat completions route

litellm/llms/bedrock/chat/chat_completions/transformation.py

@@ -22,6 +22,7 @@ import httpx
 from typing_extensions import assert_never
 
 import litellm
+from litellm.litellm_core_utils.core_helpers import set_provider_response_headers_in_hidden_params
 from litellm.litellm_core_utils.prompt_templates.image_handling import (
     async_inline_remote_media,
     inline_remote_image_urls,
@@ -383,6 +384,7 @@ class AmazonBedrockRuntimeChatCompletionsConfig(OpenAILikeChatConfig):
             api_key=api_key,
             json_mode=json_mode,
         )
+        set_provider_response_headers_in_hidden_params(response, raw_response.headers)
         for choice in response.choices:
             if not isinstance(choice, Choices) or not isinstance(choice.message.content, str):
                 continue

tests/unit/llms/bedrock/chat/chat_completions/test_bedrock_chat_completions_transformation.py

@@ -222,6 +222,18 @@ def test_completion_posts_runtime_chat_completions(local_cost_map, fake_aws_env)
     assert "inferenceConfig" not in body
 
 
+def test_completion_keeps_the_aws_request_id_as_a_provider_header(local_cost_map, fake_aws_env):
+    _, client = _recording_client(
+        json=_chat_completion_json("ok", "us.xai.grok-4.6"), headers={"x-amzn-requestid": "req-native-1"}
+    )
+    response = litellm.completion(
+        model="bedrock/chat_completions/us.xai.grok-4.6",
+        messages=[{"role": "user", "content": "hello"}],
+        client=client,
+    )
+
+    assert response._hidden_params["additional_headers"]["llm_provider-x-amzn-requestid"] == "req-native-1"
+
 def test_region_path_sends_the_bare_model_id_to_the_path_region(local_cost_map, fake_aws_env):
     requests, client = _recording_client(json=_chat_completion_json("ok", "openai.gpt-oss-20b-1:0"))
     litellm.completion(