From b450baa402527c5be47a532508d7708fd1ca9a41 Mon Sep 17 00:00:00 2001
From: yuneng <yuneng@berri.ai>
Date: Sun, 20 Sep 2026 07:47:33 +0000
Subject: [PATCH 01/13] test(llms): migrate phase 5 provider unit tests to
 tests/unit

Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
---
 .../image/test_bedrock_image_bearer_token.py  | 158 ------------------
 .../test_amazon_nova_canvas_transformation.py |   0
 .../test_amazon_stability3_transformation.py  |   0
 .../image/test_bedrock_image_bearer_token.py  |  21 +++
 .../test_bedrock_image_prepare_request.py     |   0
 .../test_amazon_nova_canvas_image_edit.py     |   0
 .../test_bedrock_agent_transformation.py      |   0
 .../guardrail_translation/test_handler.py     |   0
 ...test_bedrock_passthrough_transformation.py |   2 -
 .../realtime/test_bedrock_realtime_handler.py |   0
 .../test_bedrock_realtime_transformation.py   |   0
 .../test_bedrock_rerank_header_forwarding.py  |   0
 ...est_bedrock_vector_store_transformation.py |   0
 ...drock_mantle_passthrough_transformation.py |   0
 .../test_bfl_image_edit_transformation.py     |   0
 ...est_bfl_image_generation_transformation.py |   0
 .../test_bfl_common_utils.py                  |   0
 .../chat/test_bytez_chat_transformation.py    |   0
 .../test_cerebras_chat_transformation.py      |   0
 .../llms/chat/test_converse_handler.py        |   0
 .../chatgpt/test_chatgpt_authenticator.py     |   0
 21 files changed, 21 insertions(+), 160 deletions(-)
 delete mode 100644 tests/test_litellm/llms/bedrock/image/test_bedrock_image_bearer_token.py
 rename tests/{test_litellm => unit}/llms/bedrock/image/test_amazon_nova_canvas_transformation.py (100%)
 rename tests/{test_litellm => unit}/llms/bedrock/image/test_amazon_stability3_transformation.py (100%)
 create mode 100644 tests/unit/llms/bedrock/image/test_bedrock_image_bearer_token.py
 rename tests/{test_litellm => unit}/llms/bedrock/image/test_bedrock_image_prepare_request.py (100%)
 rename tests/{test_litellm => unit}/llms/bedrock/image_edit/test_amazon_nova_canvas_image_edit.py (100%)
 rename tests/{test_litellm => unit}/llms/bedrock/invoke_agent/test_bedrock_agent_transformation.py (100%)
 rename tests/{test_litellm => unit}/llms/bedrock/passthrough/guardrail_translation/test_handler.py (100%)
 rename tests/{test_litellm => unit}/llms/bedrock/passthrough/test_bedrock_passthrough_transformation.py (99%)
 rename tests/{test_litellm => unit}/llms/bedrock/realtime/test_bedrock_realtime_handler.py (100%)
 rename tests/{test_litellm => unit}/llms/bedrock/realtime/test_bedrock_realtime_transformation.py (100%)
 rename tests/{test_litellm => unit}/llms/bedrock/rerank/test_bedrock_rerank_header_forwarding.py (100%)
 rename tests/{test_litellm => unit}/llms/bedrock/vector_stores/test_bedrock_vector_store_transformation.py (100%)
 rename tests/{test_litellm => unit}/llms/bedrock_mantle/passthrough/test_bedrock_mantle_passthrough_transformation.py (100%)
 rename tests/{test_litellm => unit}/llms/black_forest_labs/image_edit/test_bfl_image_edit_transformation.py (100%)
 rename tests/{test_litellm => unit}/llms/black_forest_labs/image_generation/test_bfl_image_generation_transformation.py (100%)
 rename tests/{test_litellm => unit}/llms/black_forest_labs/test_bfl_common_utils.py (100%)
 rename tests/{test_litellm => unit}/llms/bytez/chat/test_bytez_chat_transformation.py (100%)
 rename tests/{test_litellm => unit}/llms/cerebras/test_cerebras_chat_transformation.py (100%)
 rename tests/{test_litellm => unit}/llms/chat/test_converse_handler.py (100%)
 rename tests/{test_litellm => unit}/llms/chatgpt/test_chatgpt_authenticator.py (100%)

diff --git a/tests/test_litellm/llms/bedrock/image/test_bedrock_image_bearer_token.py b/tests/test_litellm/llms/bedrock/image/test_bedrock_image_bearer_token.py
deleted file mode 100644
index 0b11a66c100..00000000000
--- a/tests/test_litellm/llms/bedrock/image/test_bedrock_image_bearer_token.py
+++ /dev/null
@@ -1,158 +0,0 @@
-import json
-import os
-from unittest.mock import Mock, patch
-import pytest
-
-
-import litellm
-from litellm.llms.custom_httpx.http_handler import HTTPHandler, AsyncHTTPHandler
-
-# Mock response for Bedrock image generation
-mock_image_response = {"images": ["base64_encoded_image_data"], "error": None}
-
-
-class TestBedrockImageGeneration:
-    def test_image_generation_with_api_key_bearer_token(self):
-        """Test image generation with bearer token authentication"""
-        test_api_key = "test-bearer-token-12345"
-        model = "bedrock/stability.sd3-large-v1:0"
-        prompt = "A cute baby sea otter"
-
-        with patch(
-            "litellm.llms.bedrock.image_generation.image_handler.BedrockImageGeneration.image_generation"
-        ) as mock_bedrock_image_gen:
-            # Setup mock response
-            mock_image_response_obj = litellm.ImageResponse()
-            mock_image_response_obj.data = [{"url": "https://example.com/image.jpg"}]
-            mock_bedrock_image_gen.return_value = mock_image_response_obj
-
-            response = litellm.image_generation(
-                model=model,
-                prompt=prompt,
-                aws_region_name="us-west-2",
-                api_key=test_api_key,
-            )
-
-            assert response is not None
-            assert len(response.data) > 0
-
-            mock_bedrock_image_gen.assert_called_once()
-            for call in mock_bedrock_image_gen.call_args_list:
-                if "headers" in call.kwargs:
-                    headers = call.kwargs["headers"]
-                    if (
-                        "Authorization" in headers
-                        and headers["Authorization"] == f"Bearer {test_api_key}"
-                    ):
-                        break
-
-    def test_image_generation_with_env_variable_bearer_token(self, monkeypatch):
-        """Test image generation with bearer token from environment variable"""
-        test_api_key = "env-bearer-token-12345"
-        model = "bedrock/stability.sd3-large-v1:0"
-        prompt = "A cute baby sea otter"
-
-        # Mock the environment variable
-        with (
-            patch.dict(os.environ, {"AWS_BEARER_TOKEN_BEDROCK": test_api_key}),
-            patch(
-                "litellm.llms.bedrock.image_generation.image_handler.BedrockImageGeneration.image_generation"
-            ) as mock_bedrock_image_gen,
-        ):
-
-            mock_image_response_obj = litellm.ImageResponse()
-            mock_image_response_obj.data = [{"url": "https://example.com/image.jpg"}]
-            mock_bedrock_image_gen.return_value = mock_image_response_obj
-
-            response = litellm.image_generation(
-                model=model, prompt=prompt, aws_region_name="us-west-2"
-            )
-
-            assert response is not None
-            assert len(response.data) > 0
-
-            mock_bedrock_image_gen.assert_called_once()
-            for call in mock_bedrock_image_gen.call_args_list:
-                if "headers" in call.kwargs:
-                    headers = call.kwargs["headers"]
-                    if (
-                        "Authorization" in headers
-                        and headers["Authorization"] == f"Bearer {test_api_key}"
-                    ):
-                        break
-
-    @pytest.mark.asyncio
-    async def test_async_image_generation_with_bearer_token(self):
-        """Test async image generation with bearer token authentication"""
-        test_api_key = "async-bearer-token-12345"
-        model = "bedrock/stability.sd3-large-v1:0"
-        prompt = "A cute baby sea otter"
-
-        with patch(
-            "litellm.llms.bedrock.image_generation.image_handler.BedrockImageGeneration.async_image_generation"
-        ) as mock_async_bedrock_image_gen:
-            mock_image_response_obj = litellm.ImageResponse()
-            mock_image_response_obj.data = [{"url": "https://example.com/image.jpg"}]
-            mock_async_bedrock_image_gen.return_value = mock_image_response_obj
-
-            # Call async image generation with api_key parameter
-            response = await litellm.aimage_generation(
-                model=model,
-                prompt=prompt,
-                aws_region_name="us-west-2",
-                api_key=test_api_key,
-            )
-
-            assert response is not None
-            assert len(response.data) > 0
-
-            mock_async_bedrock_image_gen.assert_called_once()
-            for call in mock_async_bedrock_image_gen.call_args_list:
-                if "headers" in call.kwargs:
-                    headers = call.kwargs["headers"]
-                    if (
-                        "Authorization" in headers
-                        and headers["Authorization"] == f"Bearer {test_api_key}"
-                    ):
-                        break
-
-    def test_image_generation_with_sigv4(self):
-        """Test image generation falls back to SigV4 auth when no bearer token is provided"""
-        model = "bedrock/stability.sd3-large-v1:0"
-        prompt = "A cute baby sea otter"
-
-        with patch(
-            "litellm.llms.bedrock.image_generation.image_handler.BedrockImageGeneration.image_generation"
-        ) as mock_bedrock_image_gen:
-            mock_image_response_obj = litellm.ImageResponse()
-            mock_image_response_obj.data = [{"url": "https://example.com/image.jpg"}]
-            mock_bedrock_image_gen.return_value = mock_image_response_obj
-
-            response = litellm.image_generation(
-                model=model, prompt=prompt, aws_region_name="us-west-2"
-            )
-
-            assert response is not None
-            assert len(response.data) > 0
-            mock_bedrock_image_gen.assert_called_once()
-
-
-def test_image_generation_bearer_token_never_runs_the_sigv4_credential_chain(monkeypatch):
-    """The deployment's AWS profile does not exist, so resolving SigV4 credentials
-    raises; a bearer-token deployment must still sign the request with the
-    bearer token alone."""
-    from litellm.llms.bedrock.image_generation.image_handler import BedrockImageGeneration
-
-    monkeypatch.setenv("AWS_BEARER_TOKEN_BEDROCK", "env-bearer-token-12345")
-
-    request = BedrockImageGeneration()._prepare_request(
-        model="amazon.nova-canvas-v1:0",
-        prompt="A cute baby sea otter",
-        optional_params={"aws_region_name": "us-west-2", "aws_profile_name": "litellm-no-such-aws-profile"},
-        api_base=None,
-        extra_headers=None,
-        api_key=None,
-        logging_obj=Mock(),
-    )
-
-    assert request.prepped.headers["Authorization"] == "Bearer env-bearer-token-12345"
diff --git a/tests/test_litellm/llms/bedrock/image/test_amazon_nova_canvas_transformation.py b/tests/unit/llms/bedrock/image/test_amazon_nova_canvas_transformation.py
similarity index 100%
rename from tests/test_litellm/llms/bedrock/image/test_amazon_nova_canvas_transformation.py
rename to tests/unit/llms/bedrock/image/test_amazon_nova_canvas_transformation.py
diff --git a/tests/test_litellm/llms/bedrock/image/test_amazon_stability3_transformation.py b/tests/unit/llms/bedrock/image/test_amazon_stability3_transformation.py
similarity index 100%
rename from tests/test_litellm/llms/bedrock/image/test_amazon_stability3_transformation.py
rename to tests/unit/llms/bedrock/image/test_amazon_stability3_transformation.py
diff --git a/tests/unit/llms/bedrock/image/test_bedrock_image_bearer_token.py b/tests/unit/llms/bedrock/image/test_bedrock_image_bearer_token.py
new file mode 100644
index 00000000000..599507da03d
--- /dev/null
+++ b/tests/unit/llms/bedrock/image/test_bedrock_image_bearer_token.py
@@ -0,0 +1,21 @@
+from unittest.mock import Mock
+
+def test_image_generation_bearer_token_never_runs_the_sigv4_credential_chain(monkeypatch):
+    """The deployment's AWS profile does not exist, so resolving SigV4 credentials
+    raises; a bearer-token deployment must still sign the request with the
+    bearer token alone."""
+    from litellm.llms.bedrock.image_generation.image_handler import BedrockImageGeneration
+
+    monkeypatch.setenv("AWS_BEARER_TOKEN_BEDROCK", "env-bearer-token-12345")
+
+    request = BedrockImageGeneration()._prepare_request(
+        model="amazon.nova-canvas-v1:0",
+        prompt="A cute baby sea otter",
+        optional_params={"aws_region_name": "us-west-2", "aws_profile_name": "litellm-no-such-aws-profile"},
+        api_base=None,
+        extra_headers=None,
+        api_key=None,
+        logging_obj=Mock(),
+    )
+
+    assert request.prepped.headers["Authorization"] == "Bearer env-bearer-token-12345"
diff --git a/tests/test_litellm/llms/bedrock/image/test_bedrock_image_prepare_request.py b/tests/unit/llms/bedrock/image/test_bedrock_image_prepare_request.py
similarity index 100%
rename from tests/test_litellm/llms/bedrock/image/test_bedrock_image_prepare_request.py
rename to tests/unit/llms/bedrock/image/test_bedrock_image_prepare_request.py
diff --git a/tests/test_litellm/llms/bedrock/image_edit/test_amazon_nova_canvas_image_edit.py b/tests/unit/llms/bedrock/image_edit/test_amazon_nova_canvas_image_edit.py
similarity index 100%
rename from tests/test_litellm/llms/bedrock/image_edit/test_amazon_nova_canvas_image_edit.py
rename to tests/unit/llms/bedrock/image_edit/test_amazon_nova_canvas_image_edit.py
diff --git a/tests/test_litellm/llms/bedrock/invoke_agent/test_bedrock_agent_transformation.py b/tests/unit/llms/bedrock/invoke_agent/test_bedrock_agent_transformation.py
similarity index 100%
rename from tests/test_litellm/llms/bedrock/invoke_agent/test_bedrock_agent_transformation.py
rename to tests/unit/llms/bedrock/invoke_agent/test_bedrock_agent_transformation.py
diff --git a/tests/test_litellm/llms/bedrock/passthrough/guardrail_translation/test_handler.py b/tests/unit/llms/bedrock/passthrough/guardrail_translation/test_handler.py
similarity index 100%
rename from tests/test_litellm/llms/bedrock/passthrough/guardrail_translation/test_handler.py
rename to tests/unit/llms/bedrock/passthrough/guardrail_translation/test_handler.py
diff --git a/tests/test_litellm/llms/bedrock/passthrough/test_bedrock_passthrough_transformation.py b/tests/unit/llms/bedrock/passthrough/test_bedrock_passthrough_transformation.py
similarity index 99%
rename from tests/test_litellm/llms/bedrock/passthrough/test_bedrock_passthrough_transformation.py
rename to tests/unit/llms/bedrock/passthrough/test_bedrock_passthrough_transformation.py
index f2a9af11af7..854ef92fa4b 100644
--- a/tests/test_litellm/llms/bedrock/passthrough/test_bedrock_passthrough_transformation.py
+++ b/tests/unit/llms/bedrock/passthrough/test_bedrock_passthrough_transformation.py
@@ -367,8 +367,6 @@ def test_bedrock_passthrough_region_extraction_from_inference_profile_arn():
         assert (
             "us-west-2" in api_base
         ), f"Expected region 'us-west-2' from ARN in base URL, but got: {api_base}"
-
-
 def test_bedrock_passthrough_model_id_arn_encoding():
     """
     Test that model_id ARNs are properly URL-encoded when used in endpoints.
diff --git a/tests/test_litellm/llms/bedrock/realtime/test_bedrock_realtime_handler.py b/tests/unit/llms/bedrock/realtime/test_bedrock_realtime_handler.py
similarity index 100%
rename from tests/test_litellm/llms/bedrock/realtime/test_bedrock_realtime_handler.py
rename to tests/unit/llms/bedrock/realtime/test_bedrock_realtime_handler.py
diff --git a/tests/test_litellm/llms/bedrock/realtime/test_bedrock_realtime_transformation.py b/tests/unit/llms/bedrock/realtime/test_bedrock_realtime_transformation.py
similarity index 100%
rename from tests/test_litellm/llms/bedrock/realtime/test_bedrock_realtime_transformation.py
rename to tests/unit/llms/bedrock/realtime/test_bedrock_realtime_transformation.py
diff --git a/tests/test_litellm/llms/bedrock/rerank/test_bedrock_rerank_header_forwarding.py b/tests/unit/llms/bedrock/rerank/test_bedrock_rerank_header_forwarding.py
similarity index 100%
rename from tests/test_litellm/llms/bedrock/rerank/test_bedrock_rerank_header_forwarding.py
rename to tests/unit/llms/bedrock/rerank/test_bedrock_rerank_header_forwarding.py
diff --git a/tests/test_litellm/llms/bedrock/vector_stores/test_bedrock_vector_store_transformation.py b/tests/unit/llms/bedrock/vector_stores/test_bedrock_vector_store_transformation.py
similarity index 100%
rename from tests/test_litellm/llms/bedrock/vector_stores/test_bedrock_vector_store_transformation.py
rename to tests/unit/llms/bedrock/vector_stores/test_bedrock_vector_store_transformation.py
diff --git a/tests/test_litellm/llms/bedrock_mantle/passthrough/test_bedrock_mantle_passthrough_transformation.py b/tests/unit/llms/bedrock_mantle/passthrough/test_bedrock_mantle_passthrough_transformation.py
similarity index 100%
rename from tests/test_litellm/llms/bedrock_mantle/passthrough/test_bedrock_mantle_passthrough_transformation.py
rename to tests/unit/llms/bedrock_mantle/passthrough/test_bedrock_mantle_passthrough_transformation.py
diff --git a/tests/test_litellm/llms/black_forest_labs/image_edit/test_bfl_image_edit_transformation.py b/tests/unit/llms/black_forest_labs/image_edit/test_bfl_image_edit_transformation.py
similarity index 100%
rename from tests/test_litellm/llms/black_forest_labs/image_edit/test_bfl_image_edit_transformation.py
rename to tests/unit/llms/black_forest_labs/image_edit/test_bfl_image_edit_transformation.py
diff --git a/tests/test_litellm/llms/black_forest_labs/image_generation/test_bfl_image_generation_transformation.py b/tests/unit/llms/black_forest_labs/image_generation/test_bfl_image_generation_transformation.py
similarity index 100%
rename from tests/test_litellm/llms/black_forest_labs/image_generation/test_bfl_image_generation_transformation.py
rename to tests/unit/llms/black_forest_labs/image_generation/test_bfl_image_generation_transformation.py
diff --git a/tests/test_litellm/llms/black_forest_labs/test_bfl_common_utils.py b/tests/unit/llms/black_forest_labs/test_bfl_common_utils.py
similarity index 100%
rename from tests/test_litellm/llms/black_forest_labs/test_bfl_common_utils.py
rename to tests/unit/llms/black_forest_labs/test_bfl_common_utils.py
diff --git a/tests/test_litellm/llms/bytez/chat/test_bytez_chat_transformation.py b/tests/unit/llms/bytez/chat/test_bytez_chat_transformation.py
similarity index 100%
rename from tests/test_litellm/llms/bytez/chat/test_bytez_chat_transformation.py
rename to tests/unit/llms/bytez/chat/test_bytez_chat_transformation.py
diff --git a/tests/test_litellm/llms/cerebras/test_cerebras_chat_transformation.py b/tests/unit/llms/cerebras/test_cerebras_chat_transformation.py
similarity index 100%
rename from tests/test_litellm/llms/cerebras/test_cerebras_chat_transformation.py
rename to tests/unit/llms/cerebras/test_cerebras_chat_transformation.py
diff --git a/tests/test_litellm/llms/chat/test_converse_handler.py b/tests/unit/llms/chat/test_converse_handler.py
similarity index 100%
rename from tests/test_litellm/llms/chat/test_converse_handler.py
rename to tests/unit/llms/chat/test_converse_handler.py
diff --git a/tests/test_litellm/llms/chatgpt/test_chatgpt_authenticator.py b/tests/unit/llms/chatgpt/test_chatgpt_authenticator.py
similarity index 100%
rename from tests/test_litellm/llms/chatgpt/test_chatgpt_authenticator.py
rename to tests/unit/llms/chatgpt/test_chatgpt_authenticator.py

From 4defed7f2e7eaacdf8e130857eddc6018c7bc3f7 Mon Sep 17 00:00:00 2001
From: yuneng <yuneng@berri.ai>
Date: Sun, 20 Sep 2026 08:03:54 +0000
Subject: [PATCH 02/13] test: migrate wave 1 phase 8 legacy llm tests to
 tests/unit

Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
---
 .../chat/test_hosted_vllm_ssl_verify.py       | 147 ------------------
 .../test_hosted_vllm_embedding_ssl_verify.py  | 135 ----------------
 ..._github_copilot_messages_transformation.py |   7 -
 ...github_copilot_responses_transformation.py | 116 +++++---------
 .../test_gradient_ai_chat_transformation.py   |   0
 .../chat/test_groq_chat_transformation.py     |   2 -
 .../llms/groq/test_groq_cost_calculator.py    |   0
 .../test_hosted_vllm_chat_transformation.py   |  71 +--------
 ...st_hosted_vllm_embedding_transformation.py |   8 +-
 ...t_hosted_vllm_image_edit_transformation.py |   0
 .../responses/test_hosted_vllm_responses.py   |   9 +-
 .../test_hosted_vllm_rerank_transformation.py |   0
 .../test_hosted_vllm_video_transformation.py  |   0
 .../test_huggingface_rerank_transformation.py |  40 +----
 .../test_inception_chat_transformation.py     |  18 +--
 ...est_inception_completion_transformation.py |  18 +--
 .../test_jina_embedding_transformation.py     |   0
 .../chat/test_langflow_chat_transformation.py |  31 +---
 .../litellm_proxy/test_sandbox_executor.py    |  25 +--
 .../litellm_proxy/test_skills_ownership.py    |  73 ++-------
 20 files changed, 84 insertions(+), 616 deletions(-)
 delete mode 100644 tests/test_litellm/llms/hosted_vllm/chat/test_hosted_vllm_ssl_verify.py
 delete mode 100644 tests/test_litellm/llms/hosted_vllm/embedding/test_hosted_vllm_embedding_ssl_verify.py
 rename tests/{test_litellm => unit}/llms/github_copilot/messages/test_github_copilot_messages_transformation.py (98%)
 rename tests/{test_litellm => unit}/llms/github_copilot/responses/test_github_copilot_responses_transformation.py (89%)
 rename tests/{test_litellm => unit}/llms/gradient_ai/chat/test_gradient_ai_chat_transformation.py (100%)
 rename tests/{test_litellm => unit}/llms/groq/chat/test_groq_chat_transformation.py (99%)
 rename tests/{test_litellm => unit}/llms/groq/test_groq_cost_calculator.py (100%)
 rename tests/{test_litellm => unit}/llms/hosted_vllm/chat/test_hosted_vllm_chat_transformation.py (82%)
 rename tests/{test_litellm => unit}/llms/hosted_vllm/embedding/test_hosted_vllm_embedding_transformation.py (97%)
 rename tests/{test_litellm => unit}/llms/hosted_vllm/image_edit/test_hosted_vllm_image_edit_transformation.py (100%)
 rename tests/{test_litellm => unit}/llms/hosted_vllm/responses/test_hosted_vllm_responses.py (96%)
 rename tests/{test_litellm => unit}/llms/hosted_vllm/test_hosted_vllm_rerank_transformation.py (100%)
 rename tests/{test_litellm => unit}/llms/hosted_vllm/videos/test_hosted_vllm_video_transformation.py (100%)
 rename tests/{test_litellm => unit}/llms/huggingface/rerank/test_huggingface_rerank_transformation.py (91%)
 rename tests/{test_litellm => unit}/llms/inception/test_inception_chat_transformation.py (96%)
 rename tests/{test_litellm => unit}/llms/inception/test_inception_completion_transformation.py (95%)
 rename tests/{test_litellm => unit}/llms/jina_ai/embedding/test_jina_embedding_transformation.py (100%)
 rename tests/{test_litellm => unit}/llms/langflow/chat/test_langflow_chat_transformation.py (93%)
 rename tests/{test_litellm => unit}/llms/litellm_proxy/test_sandbox_executor.py (84%)
 rename tests/{test_litellm => unit}/llms/litellm_proxy/test_skills_ownership.py (88%)

diff --git a/tests/test_litellm/llms/hosted_vllm/chat/test_hosted_vllm_ssl_verify.py b/tests/test_litellm/llms/hosted_vllm/chat/test_hosted_vllm_ssl_verify.py
deleted file mode 100644
index 2364468efe1..00000000000
--- a/tests/test_litellm/llms/hosted_vllm/chat/test_hosted_vllm_ssl_verify.py
+++ /dev/null
@@ -1,147 +0,0 @@
-"""
-Test SSL verification for hosted_vllm provider.
-
-This test ensures that the ssl_verify parameter is properly passed through
-to the HTTP client when using the hosted_vllm provider.
-
-Issue: ssl_verify parameter was being ignored because hosted_vllm fell through
-to the OpenAI catch-all path in main.py, which doesn't pass ssl_verify to the HTTP client.
-"""
-
-from unittest.mock import MagicMock, patch
-
-import pytest
-
-
-import litellm
-
-
-class TestHostedVLLMSSLVerify:
-    """Test suite for SSL verification in hosted_vllm provider."""
-
-    @patch("litellm.llms.custom_httpx.llm_http_handler._get_httpx_client")
-    def test_hosted_vllm_ssl_verify_false_sync(self, mock_get_httpx_client):
-        """Test that ssl_verify=False is passed to the HTTP client for sync calls."""
-        # Setup mock client
-        mock_client = MagicMock()
-        mock_response = MagicMock()
-        mock_response.status_code = 200
-        mock_response.headers = {"content-type": "application/json"}
-        mock_response.json.return_value = {
-            "id": "chatcmpl-test",
-            "object": "chat.completion",
-            "created": 1234567890,
-            "model": "test-model",
-            "choices": [
-                {
-                    "index": 0,
-                    "message": {
-                        "role": "assistant",
-                        "content": "Test response",
-                    },
-                    "finish_reason": "stop",
-                }
-            ],
-            "usage": {
-                "prompt_tokens": 10,
-                "completion_tokens": 5,
-                "total_tokens": 15,
-            },
-        }
-        mock_response.text = '{"id": "chatcmpl-test", "object": "chat.completion", "created": 1234567890, "model": "test-model", "choices": [{"index": 0, "message": {"role": "assistant", "content": "Test response"}, "finish_reason": "stop"}], "usage": {"prompt_tokens": 10, "completion_tokens": 5, "total_tokens": 15}}'
-        mock_client.post.return_value = mock_response
-        mock_get_httpx_client.return_value = mock_client
-
-        try:
-            litellm.completion(
-                model="hosted_vllm/test-model",
-                messages=[{"role": "user", "content": "Hello"}],
-                api_base="https://test-vllm.example.com/v1",
-                ssl_verify=False,
-            )
-        except Exception:
-            # Even if the response parsing fails, we just need to verify
-            # that the mock was called with the correct ssl_verify parameter
-            pass
-
-        # Verify _get_httpx_client was called with ssl_verify=False
-        mock_get_httpx_client.assert_called()
-        call_args = mock_get_httpx_client.call_args
-
-        # Check that params contains ssl_verify=False
-        if call_args[0]:
-            # Positional argument
-            params = call_args[0][0]
-        else:
-            # Keyword argument
-            params = call_args[1].get("params", {})
-
-        assert (
-            params.get("ssl_verify") is False
-        ), f"Expected ssl_verify=False in params, got {params}"
-
-    @patch("litellm.llms.custom_httpx.llm_http_handler.get_async_httpx_client")
-    @pytest.mark.asyncio
-    async def test_hosted_vllm_ssl_verify_false_async(
-        self, mock_get_async_httpx_client
-    ):
-        """Test that ssl_verify=False is passed to the HTTP client for async calls."""
-        # Setup mock async client
-        mock_client = MagicMock()
-        mock_response = MagicMock()
-        mock_response.status_code = 200
-        mock_response.headers = {"content-type": "application/json"}
-        mock_response.json.return_value = {
-            "id": "chatcmpl-test",
-            "object": "chat.completion",
-            "created": 1234567890,
-            "model": "test-model",
-            "choices": [
-                {
-                    "index": 0,
-                    "message": {
-                        "role": "assistant",
-                        "content": "Test response",
-                    },
-                    "finish_reason": "stop",
-                }
-            ],
-            "usage": {
-                "prompt_tokens": 10,
-                "completion_tokens": 5,
-                "total_tokens": 15,
-            },
-        }
-        mock_response.text = '{"id": "chatcmpl-test", "object": "chat.completion", "created": 1234567890, "model": "test-model", "choices": [{"index": 0, "message": {"role": "assistant", "content": "Test response"}, "finish_reason": "stop"}], "usage": {"prompt_tokens": 10, "completion_tokens": 5, "total_tokens": 15}}'
-
-        async def mock_post(*args, **kwargs):
-            return mock_response
-
-        mock_client.post = mock_post
-        mock_get_async_httpx_client.return_value = mock_client
-
-        try:
-            await litellm.acompletion(
-                model="hosted_vllm/test-model",
-                messages=[{"role": "user", "content": "Hello"}],
-                api_base="https://test-vllm.example.com/v1",
-                ssl_verify=False,
-            )
-        except Exception:
-            # Even if the response parsing fails, we just need to verify
-            # that the mock was called with the correct ssl_verify parameter
-            pass
-
-        # Verify get_async_httpx_client was called with ssl_verify=False
-        mock_get_async_httpx_client.assert_called()
-        call_kwargs = mock_get_async_httpx_client.call_args[1]
-
-        # Check that params contains ssl_verify=False
-        params = call_kwargs.get("params", {})
-        assert (
-            params.get("ssl_verify") is False
-        ), f"Expected ssl_verify=False in params, got {params}"
-
-
-if __name__ == "__main__":
-    pytest.main([__file__, "-v", "-s"])
diff --git a/tests/test_litellm/llms/hosted_vllm/embedding/test_hosted_vllm_embedding_ssl_verify.py b/tests/test_litellm/llms/hosted_vllm/embedding/test_hosted_vllm_embedding_ssl_verify.py
deleted file mode 100644
index de94da49384..00000000000
--- a/tests/test_litellm/llms/hosted_vllm/embedding/test_hosted_vllm_embedding_ssl_verify.py
+++ /dev/null
@@ -1,135 +0,0 @@
-"""
-Test SSL verification for hosted_vllm provider embeddings.
-
-This test ensures that the ssl_verify parameter is properly passed through
-to the HTTP client when using the hosted_vllm provider for embeddings.
-
-Issue: ssl_verify parameter was being ignored because hosted_vllm fell through
-to the openai_like catch-all path in main.py, which doesn't pass ssl_verify to the HTTP client.
-"""
-
-from unittest.mock import MagicMock, patch
-
-import pytest
-
-
-import litellm
-
-
-class TestHostedVLLMEmbeddingSSLVerify:
-    """Test suite for SSL verification in hosted_vllm provider embeddings."""
-
-    @patch("litellm.llms.custom_httpx.llm_http_handler._get_httpx_client")
-    def test_hosted_vllm_embedding_ssl_verify_false_sync(self, mock_get_httpx_client):
-        """Test that ssl_verify=False is passed to the HTTP client for sync embedding calls."""
-        # Setup mock client
-        mock_client = MagicMock()
-        mock_response = MagicMock()
-        mock_response.status_code = 200
-        mock_response.headers = {"content-type": "application/json"}
-        mock_response.json.return_value = {
-            "object": "list",
-            "data": [
-                {
-                    "object": "embedding",
-                    "index": 0,
-                    "embedding": [0.1, 0.2, 0.3, 0.4, 0.5],
-                }
-            ],
-            "model": "text-embedding-model",
-            "usage": {
-                "prompt_tokens": 5,
-                "total_tokens": 5,
-            },
-        }
-        mock_response.text = '{"object": "list", "data": [{"object": "embedding", "index": 0, "embedding": [0.1, 0.2, 0.3, 0.4, 0.5]}], "model": "text-embedding-model", "usage": {"prompt_tokens": 5, "total_tokens": 5}}'
-        mock_client.post.return_value = mock_response
-        mock_get_httpx_client.return_value = mock_client
-
-        try:
-            litellm.embedding(
-                model="hosted_vllm/text-embedding-model",
-                input=["hello world"],
-                api_base="https://test-vllm.example.com/v1",
-                ssl_verify=False,
-            )
-        except Exception:
-            # Even if the response parsing fails, we just need to verify
-            # that the mock was called with the correct ssl_verify parameter
-            pass
-
-        # Verify _get_httpx_client was called with ssl_verify=False
-        mock_get_httpx_client.assert_called()
-        call_args = mock_get_httpx_client.call_args
-
-        # Check that params contains ssl_verify=False
-        if call_args[0]:
-            # Positional argument
-            params = call_args[0][0]
-        else:
-            # Keyword argument
-            params = call_args[1].get("params", {})
-
-        assert (
-            params.get("ssl_verify") is False
-        ), f"Expected ssl_verify=False in params, got {params}"
-
-    @patch("litellm.llms.custom_httpx.llm_http_handler.get_async_httpx_client")
-    @pytest.mark.asyncio
-    async def test_hosted_vllm_embedding_ssl_verify_false_async(
-        self, mock_get_async_httpx_client
-    ):
-        """Test that ssl_verify=False is passed to the HTTP client for async embedding calls."""
-        # Setup mock async client
-        mock_client = MagicMock()
-        mock_response = MagicMock()
-        mock_response.status_code = 200
-        mock_response.headers = {"content-type": "application/json"}
-        mock_response.json.return_value = {
-            "object": "list",
-            "data": [
-                {
-                    "object": "embedding",
-                    "index": 0,
-                    "embedding": [0.1, 0.2, 0.3, 0.4, 0.5],
-                }
-            ],
-            "model": "text-embedding-model",
-            "usage": {
-                "prompt_tokens": 5,
-                "total_tokens": 5,
-            },
-        }
-        mock_response.text = '{"object": "list", "data": [{"object": "embedding", "index": 0, "embedding": [0.1, 0.2, 0.3, 0.4, 0.5]}], "model": "text-embedding-model", "usage": {"prompt_tokens": 5, "total_tokens": 5}}'
-
-        async def mock_post(*args, **kwargs):
-            return mock_response
-
-        mock_client.post = mock_post
-        mock_get_async_httpx_client.return_value = mock_client
-
-        try:
-            await litellm.aembedding(
-                model="hosted_vllm/text-embedding-model",
-                input=["hello world"],
-                api_base="https://test-vllm.example.com/v1",
-                ssl_verify=False,
-            )
-        except Exception:
-            # Even if the response parsing fails, we just need to verify
-            # that the mock was called with the correct ssl_verify parameter
-            pass
-
-        # Verify get_async_httpx_client was called with ssl_verify=False
-        mock_get_async_httpx_client.assert_called()
-        call_kwargs = mock_get_async_httpx_client.call_args[1]
-
-        # Check that params contains ssl_verify=False
-        params = call_kwargs.get("params", {})
-        assert (
-            params.get("ssl_verify") is False
-        ), f"Expected ssl_verify=False in params, got {params}"
-
-
-if __name__ == "__main__":
-    pytest.main([__file__, "-v", "-s"])
diff --git a/tests/test_litellm/llms/github_copilot/messages/test_github_copilot_messages_transformation.py b/tests/unit/llms/github_copilot/messages/test_github_copilot_messages_transformation.py
similarity index 98%
rename from tests/test_litellm/llms/github_copilot/messages/test_github_copilot_messages_transformation.py
rename to tests/unit/llms/github_copilot/messages/test_github_copilot_messages_transformation.py
index 8039e744f46..9e9760650cf 100644
--- a/tests/test_litellm/llms/github_copilot/messages/test_github_copilot_messages_transformation.py
+++ b/tests/unit/llms/github_copilot/messages/test_github_copilot_messages_transformation.py
@@ -10,13 +10,6 @@ from litellm.llms.github_copilot.messages.transformation import (
 )
 
 
-def test_github_copilot_anthropic_messages_config_init():
-    """Test GithubCopilotAnthropicMessagesConfig initialization."""
-    config = GithubCopilotAnthropicMessagesConfig()
-    assert config is not None
-    assert hasattr(config, "authenticator")
-
-
 def test_github_copilot_anthropic_messages_get_complete_url():
     """get_complete_url builds the /v1/messages URL from the base it is handed.
 
diff --git a/tests/test_litellm/llms/github_copilot/responses/test_github_copilot_responses_transformation.py b/tests/unit/llms/github_copilot/responses/test_github_copilot_responses_transformation.py
similarity index 89%
rename from tests/test_litellm/llms/github_copilot/responses/test_github_copilot_responses_transformation.py
rename to tests/unit/llms/github_copilot/responses/test_github_copilot_responses_transformation.py
index 0174465b0cc..b8380b7adb4 100644
--- a/tests/test_litellm/llms/github_copilot/responses/test_github_copilot_responses_transformation.py
+++ b/tests/unit/llms/github_copilot/responses/test_github_copilot_responses_transformation.py
@@ -26,9 +26,7 @@ def use_local_model_cost_map(monkeypatch: pytest.MonkeyPatch):
     """Pin litellm.model_cost to the bundled local backup so tests don't depend
     on remote catalog fetches (and don't change behavior across remote refreshes)."""
     monkeypatch.setenv("LITELLM_LOCAL_MODEL_COST_MAP", "True")
-    monkeypatch.setattr(
-        litellm, "model_cost", get_model_cost_map(url=litellm.model_cost_map_url)
-    )
+    monkeypatch.setattr(litellm, "model_cost", get_model_cost_map(url=litellm.model_cost_map_url))
     litellm.add_known_models(model_cost_map=litellm.model_cost)
 
 
@@ -44,49 +42,35 @@ class TestGithubCopilotResponsesAPITransformation:
             provider=LlmProviders.GITHUB_COPILOT,
         )
 
-        assert (
-            config is not None
-        ), "Config should not be None for GitHub Copilot provider"
-        assert isinstance(
-            config, GithubCopilotResponsesAPIConfig
-        ), f"Expected GithubCopilotResponsesAPIConfig, got {type(config)}"
-        assert (
-            config.custom_llm_provider == LlmProviders.GITHUB_COPILOT
-        ), "custom_llm_provider should be GITHUB_COPILOT"
+        assert config is not None, "Config should not be None for GitHub Copilot provider"
+        assert isinstance(config, GithubCopilotResponsesAPIConfig), (
+            f"Expected GithubCopilotResponsesAPIConfig, got {type(config)}"
+        )
+        assert config.custom_llm_provider == LlmProviders.GITHUB_COPILOT, "custom_llm_provider should be GITHUB_COPILOT"
 
     @patch("litellm.llms.github_copilot.responses.transformation.Authenticator")
     def test_github_copilot_responses_endpoint_url(self, mock_authenticator_class):
         """Test that get_complete_url returns correct GitHub Copilot endpoint"""
         # Mock authenticator to return default base
         mock_auth_instance = MagicMock()
-        mock_auth_instance.get_api_base.return_value = (
-            "https://api.individual.githubcopilot.com"
-        )
+        mock_auth_instance.get_api_base.return_value = "https://api.individual.githubcopilot.com"
         mock_authenticator_class.return_value = mock_auth_instance
 
         config = GithubCopilotResponsesAPIConfig()
 
         # Test with default GitHub Copilot API base (from authenticator)
         url = config.get_complete_url(api_base=None, litellm_params={})
-        assert (
-            url == "https://api.individual.githubcopilot.com/responses"
-        ), f"Expected GitHub Copilot responses endpoint, got {url}"
+        assert url == "https://api.individual.githubcopilot.com/responses", (
+            f"Expected GitHub Copilot responses endpoint, got {url}"
+        )
 
         # Test with custom api_base (overrides authenticator)
-        custom_url = config.get_complete_url(
-            api_base="https://custom.githubcopilot.com", litellm_params={}
-        )
-        assert (
-            custom_url == "https://custom.githubcopilot.com/responses"
-        ), f"Expected custom endpoint, got {custom_url}"
+        custom_url = config.get_complete_url(api_base="https://custom.githubcopilot.com", litellm_params={})
+        assert custom_url == "https://custom.githubcopilot.com/responses", f"Expected custom endpoint, got {custom_url}"
 
         # Test with trailing slash
-        url_with_slash = config.get_complete_url(
-            api_base="https://api.githubcopilot.com/", litellm_params={}
-        )
-        assert (
-            url_with_slash == "https://api.githubcopilot.com/responses"
-        ), "Should handle trailing slash"
+        url_with_slash = config.get_complete_url(api_base="https://api.githubcopilot.com/", litellm_params={})
+        assert url_with_slash == "https://api.githubcopilot.com/responses", "Should handle trailing slash"
 
     @patch("litellm.llms.github_copilot.responses.transformation.Authenticator")
     def test_validate_environment_default_headers(self, mock_authenticator_class):
@@ -98,9 +82,7 @@ class TestGithubCopilotResponsesAPITransformation:
 
         config = GithubCopilotResponsesAPIConfig()
 
-        headers = config.validate_environment(
-            headers={}, model="gpt-5.1-codex", litellm_params={}
-        )
+        headers = config.validate_environment(headers={}, model="gpt-5.1-codex", litellm_params={})
 
         # Check required headers
         assert headers["Authorization"] == "Bearer test-api-key-123"
@@ -127,9 +109,7 @@ class TestGithubCopilotResponsesAPITransformation:
             "custom-header": "custom-value",
         }
 
-        headers = config.validate_environment(
-            headers=custom_headers, model="gpt-5.1-codex", litellm_params={}
-        )
+        headers = config.validate_environment(headers=custom_headers, model="gpt-5.1-codex", litellm_params={})
 
         # User header should override default
         assert headers["editor-version"] == "custom/2.0.0"
@@ -182,9 +162,7 @@ class TestGithubCopilotResponsesAPITransformation:
         """Test _has_vision_input detects input_image type"""
         config = GithubCopilotResponsesAPIConfig()
 
-        input_with_vision = [
-            {"role": "user", "content": [{"type": "input_image", "data": "base64..."}]}
-        ]
+        input_with_vision = [{"role": "user", "content": [{"type": "input_image", "data": "base64..."}]}]
 
         has_vision = config._has_vision_input(input_with_vision)
         assert has_vision is True, "Should detect input_image type"
@@ -246,13 +224,11 @@ class TestGithubCopilotResponsesAPITransformation:
             }
         ]
 
-        headers = config.validate_environment(
-            headers={}, model="gpt-5.1-codex", litellm_params=mock_litellm_params
-        )
+        headers = config.validate_environment(headers={}, model="gpt-5.1-codex", litellm_params=mock_litellm_params)
 
-        assert (
-            headers.get("copilot-vision-request") == "true"
-        ), "Should add copilot-vision-request header for vision input"
+        assert headers.get("copilot-vision-request") == "true", (
+            "Should add copilot-vision-request header for vision input"
+        )
 
     @patch("litellm.llms.github_copilot.responses.transformation.Authenticator")
     def test_validate_environment_with_x_initiator(self, mock_authenticator_class):
@@ -270,21 +246,15 @@ class TestGithubCopilotResponsesAPITransformation:
             {"role": "assistant", "content": "Hi"},
         ]
 
-        headers = config.validate_environment(
-            headers={}, model="gpt-5.1-codex", litellm_params=mock_litellm_params
-        )
+        headers = config.validate_environment(headers={}, model="gpt-5.1-codex", litellm_params=mock_litellm_params)
 
-        assert (
-            headers.get("X-Initiator") == "agent"
-        ), "Should set X-Initiator to 'agent' for assistant role"
+        assert headers.get("X-Initiator") == "agent", "Should set X-Initiator to 'agent' for assistant role"
 
     def test_map_openai_params_no_transformation(self):
         """Test that map_openai_params passes through parameters unchanged"""
         config = GithubCopilotResponsesAPIConfig()
 
-        params = ResponsesAPIOptionalRequestParams(
-            temperature=0.7, max_output_tokens=1000, stream=False
-        )
+        params = ResponsesAPIOptionalRequestParams(temperature=0.7, max_output_tokens=1000, stream=False)
 
         result = config.map_openai_params(
             response_api_optional_params=params,
@@ -338,9 +308,9 @@ class TestGithubCopilotResponsesAPITransformation:
         result = config._handle_reasoning_item(reasoning_item)
 
         # encrypted_content should be preserved
-        assert (
-            result.get("encrypted_content") == "encrypted-blob-abc123"
-        ), "encrypted_content must be preserved for GitHub Copilot multi-turn conversations"
+        assert result.get("encrypted_content") == "encrypted-blob-abc123", (
+            "encrypted_content must be preserved for GitHub Copilot multi-turn conversations"
+        )
         # status=None should be filtered out
         assert "status" not in result, "status=None should be filtered out"
         # content=None should be filtered out
@@ -393,9 +363,7 @@ class TestGithubCopilotResponsesAPIRouting:
     in the (already-merged) model info; otherwise returns None so the dispatcher
     routes through the chat-completions translation bridge."""
 
-    @patch(
-        "litellm.llms.github_copilot.responses.transformation._cached_get_model_info_helper"
-    )
+    @patch("litellm.llms.github_copilot.responses.transformation._cached_get_model_info_helper")
     def test_returns_config_when_mode_is_responses(self, mock_get_info):
         """``mode=responses`` returns native config."""
         mock_get_info.return_value = {"mode": "responses"}
@@ -405,9 +373,7 @@ class TestGithubCopilotResponsesAPIRouting:
         )
         assert isinstance(config, GithubCopilotResponsesAPIConfig)
 
-    @patch(
-        "litellm.llms.github_copilot.responses.transformation._cached_get_model_info_helper"
-    )
+    @patch("litellm.llms.github_copilot.responses.transformation._cached_get_model_info_helper")
     def test_returns_none_when_mode_is_chat(self, mock_get_info):
         """``mode=chat`` returns None so dispatcher uses bridge."""
         mock_get_info.return_value = {"mode": "chat"}
@@ -417,9 +383,7 @@ class TestGithubCopilotResponsesAPIRouting:
         )
         assert config is None
 
-    @patch(
-        "litellm.llms.github_copilot.responses.transformation._cached_get_model_info_helper"
-    )
+    @patch("litellm.llms.github_copilot.responses.transformation._cached_get_model_info_helper")
     def test_returns_none_when_mode_is_unset_and_no_endpoints(self, mock_get_info):
         """Entry without ``mode`` and without ``supported_endpoints`` returns None
         (conservative default)."""
@@ -499,9 +463,7 @@ class TestGithubCopilotResponsesAPIRouting:
         )
         assert isinstance(config, GithubCopilotResponsesAPIConfig)
 
-    @patch(
-        "litellm.llms.github_copilot.responses.transformation._cached_get_model_info_helper"
-    )
+    @patch("litellm.llms.github_copilot.responses.transformation._cached_get_model_info_helper")
     def test_returns_none_when_get_model_info_raises(self, mock_get_info):
         """Catalog lookup failure (model not registered) returns None
         (conservative default; bridge handles unknown models safely)."""
@@ -512,9 +474,7 @@ class TestGithubCopilotResponsesAPIRouting:
         )
         assert config is None
 
-    @patch(
-        "litellm.llms.github_copilot.responses.transformation._cached_get_model_info_helper"
-    )
+    @patch("litellm.llms.github_copilot.responses.transformation._cached_get_model_info_helper")
     def test_user_override_via_register_model(self, mock_get_info):
         """User-supplied per-deployment ``model_info`` flows through
         ``litellm.register_model`` (called by the router) into the merged
@@ -528,9 +488,7 @@ class TestGithubCopilotResponsesAPIRouting:
         )
         assert isinstance(config, GithubCopilotResponsesAPIConfig)
 
-    @patch(
-        "litellm.llms.github_copilot.responses.transformation._cached_get_model_info_helper"
-    )
+    @patch("litellm.llms.github_copilot.responses.transformation._cached_get_model_info_helper")
     def test_realistic_chat_only_entry_returns_none(self, mock_get_info):
         """Realistic ``model_prices_and_context_window.json`` shape for a
         chat-only Copilot model (e.g. github_copilot/gemini-3.1-pro-preview)
@@ -554,9 +512,7 @@ class TestGithubCopilotResponsesAPIRouting:
         )
         assert config is None
 
-    @patch(
-        "litellm.llms.github_copilot.responses.transformation._cached_get_model_info_helper"
-    )
+    @patch("litellm.llms.github_copilot.responses.transformation._cached_get_model_info_helper")
     def test_realistic_responses_only_entry_returns_config(self, mock_get_info):
         """Realistic catalog entry for a Responses-only Copilot model
         (e.g. github_copilot/gpt-5.5) returns the native config."""
@@ -592,9 +548,7 @@ class TestGithubCopilotReasoningStreamItemIdNormalization:
     output_index group to the id from its output_item.added."""
 
     def _config(self):
-        with patch(
-            "litellm.llms.github_copilot.responses.transformation.Authenticator"
-        ):
+        with patch("litellm.llms.github_copilot.responses.transformation.Authenticator"):
             return GithubCopilotResponsesAPIConfig()
 
     def _transform(self, config, chunk):
diff --git a/tests/test_litellm/llms/gradient_ai/chat/test_gradient_ai_chat_transformation.py b/tests/unit/llms/gradient_ai/chat/test_gradient_ai_chat_transformation.py
similarity index 100%
rename from tests/test_litellm/llms/gradient_ai/chat/test_gradient_ai_chat_transformation.py
rename to tests/unit/llms/gradient_ai/chat/test_gradient_ai_chat_transformation.py
diff --git a/tests/test_litellm/llms/groq/chat/test_groq_chat_transformation.py b/tests/unit/llms/groq/chat/test_groq_chat_transformation.py
similarity index 99%
rename from tests/test_litellm/llms/groq/chat/test_groq_chat_transformation.py
rename to tests/unit/llms/groq/chat/test_groq_chat_transformation.py
index f605958b979..f5a7a920124 100644
--- a/tests/test_litellm/llms/groq/chat/test_groq_chat_transformation.py
+++ b/tests/unit/llms/groq/chat/test_groq_chat_transformation.py
@@ -202,5 +202,3 @@ class TestGroqWebSearchUsageSignal:
         model_response = litellm.ModelResponse()
         GroqChatConfig()._add_web_search_usage(model_response=model_response)
         assert getattr(model_response, "usage", None) is None
-
-
diff --git a/tests/test_litellm/llms/groq/test_groq_cost_calculator.py b/tests/unit/llms/groq/test_groq_cost_calculator.py
similarity index 100%
rename from tests/test_litellm/llms/groq/test_groq_cost_calculator.py
rename to tests/unit/llms/groq/test_groq_cost_calculator.py
diff --git a/tests/test_litellm/llms/hosted_vllm/chat/test_hosted_vllm_chat_transformation.py b/tests/unit/llms/hosted_vllm/chat/test_hosted_vllm_chat_transformation.py
similarity index 82%
rename from tests/test_litellm/llms/hosted_vllm/chat/test_hosted_vllm_chat_transformation.py
rename to tests/unit/llms/hosted_vllm/chat/test_hosted_vllm_chat_transformation.py
index 82b05601a85..1cc6a1457fc 100644
--- a/tests/test_litellm/llms/hosted_vllm/chat/test_hosted_vllm_chat_transformation.py
+++ b/tests/unit/llms/hosted_vllm/chat/test_hosted_vllm_chat_transformation.py
@@ -41,74 +41,9 @@ def test_hosted_vllm_chat_transformation_file_url():
     ]
 
 
-def test_hosted_vllm_chat_transformation_with_audio_url():
-    from litellm import completion
-
-    mock_client = MagicMock()
-    mock_response = MagicMock()
-    mock_response.status_code = 200
-    mock_response.headers = {"content-type": "application/json"}
-    mock_response.json.return_value = {
-        "id": "chatcmpl-test",
-        "object": "chat.completion",
-        "created": 1234567890,
-        "model": "llama-3.1-70b-instruct",
-        "choices": [
-            {
-                "index": 0,
-                "message": {"role": "assistant", "content": "Test response"},
-                "finish_reason": "stop",
-            }
-        ],
-        "usage": {"prompt_tokens": 10, "completion_tokens": 5, "total_tokens": 15},
-    }
-    mock_response.text = json.dumps(mock_response.json.return_value)
-    mock_client.post.return_value = mock_response
-
-    with patch(
-        "litellm.llms.custom_httpx.llm_http_handler._get_httpx_client",
-        return_value=mock_client,
-    ):
-        try:
-            completion(
-                model="hosted_vllm/llama-3.1-70b-instruct",
-                messages=[
-                    {
-                        "role": "user",
-                        "content": [
-                            {
-                                "type": "audio_url",
-                                "audio_url": {"url": "https://example.com/audio.mp3"},
-                            },
-                        ],
-                    },
-                ],
-                api_base="https://test-vllm.example.com/v1",
-            )
-        except Exception:
-            pass
-
-        mock_client.post.assert_called_once()
-        call_kwargs = mock_client.post.call_args[1]
-        request_data = json.loads(call_kwargs["data"])
-        assert request_data["messages"] == [
-            {
-                "role": "user",
-                "content": [
-                    {
-                        "type": "audio_url",
-                        "audio_url": {"url": "https://example.com/audio.mp3"},
-                    }
-                ],
-            }
-        ]
-
-
 def test_hosted_vllm_supports_reasoning_effort():
     config = HostedVLLMChatConfig()
-    supported_params = config.get_supported_openai_params(
-        model="hosted_vllm/gpt-oss-120b"
-    )
+    supported_params = config.get_supported_openai_params(model="hosted_vllm/gpt-oss-120b")
     assert "reasoning_effort" in supported_params
     optional_params = config.map_openai_params(
         non_default_params={"reasoning_effort": "high"},
@@ -129,9 +64,7 @@ def test_hosted_vllm_supports_thinking():
     Related issue: https://github.com/BerriAI/litellm/issues/19761
     """
     config = HostedVLLMChatConfig()
-    supported_params = config.get_supported_openai_params(
-        model="hosted_vllm/GLM-4.6-FP8"
-    )
+    supported_params = config.get_supported_openai_params(model="hosted_vllm/GLM-4.6-FP8")
     assert "thinking" in supported_params
 
     # Test thinking below the low threshold -> "minimal"
diff --git a/tests/test_litellm/llms/hosted_vllm/embedding/test_hosted_vllm_embedding_transformation.py b/tests/unit/llms/hosted_vllm/embedding/test_hosted_vllm_embedding_transformation.py
similarity index 97%
rename from tests/test_litellm/llms/hosted_vllm/embedding/test_hosted_vllm_embedding_transformation.py
rename to tests/unit/llms/hosted_vllm/embedding/test_hosted_vllm_embedding_transformation.py
index 34be3e12abd..5854b1596b4 100644
--- a/tests/test_litellm/llms/hosted_vllm/embedding/test_hosted_vllm_embedding_transformation.py
+++ b/tests/unit/llms/hosted_vllm/embedding/test_hosted_vllm_embedding_transformation.py
@@ -87,9 +87,7 @@ class TestHostedVLLMEmbeddingTransformation:
             headers={},
         )
 
-        assert (
-            "encoding_format" not in result
-        ), "encoding_format should not be in request when not provided"
+        assert "encoding_format" not in result, "encoding_format should not be in request when not provided"
 
     def test_encoding_format_not_included_when_none(self):
         """
@@ -278,9 +276,7 @@ class TestHostedVLLMEmbeddingTransformation:
             sent_data = json.loads(call_kwargs["data"])
 
             # Assert that encoding_format is NOT in the sent data
-            assert (
-                "encoding_format" not in sent_data
-            ), "encoding_format should not be in request when not provided"
+            assert "encoding_format" not in sent_data, "encoding_format should not be in request when not provided"
             assert sent_data["model"] == "BAAI/bge-small-en-v1.5"
             assert sent_data["input"] == ["Hello world"]
 
diff --git a/tests/test_litellm/llms/hosted_vllm/image_edit/test_hosted_vllm_image_edit_transformation.py b/tests/unit/llms/hosted_vllm/image_edit/test_hosted_vllm_image_edit_transformation.py
similarity index 100%
rename from tests/test_litellm/llms/hosted_vllm/image_edit/test_hosted_vllm_image_edit_transformation.py
rename to tests/unit/llms/hosted_vllm/image_edit/test_hosted_vllm_image_edit_transformation.py
diff --git a/tests/test_litellm/llms/hosted_vllm/responses/test_hosted_vllm_responses.py b/tests/unit/llms/hosted_vllm/responses/test_hosted_vllm_responses.py
similarity index 96%
rename from tests/test_litellm/llms/hosted_vllm/responses/test_hosted_vllm_responses.py
rename to tests/unit/llms/hosted_vllm/responses/test_hosted_vllm_responses.py
index e81bf0c4f1f..55d0ce1e68e 100644
--- a/tests/test_litellm/llms/hosted_vllm/responses/test_hosted_vllm_responses.py
+++ b/tests/unit/llms/hosted_vllm/responses/test_hosted_vllm_responses.py
@@ -68,9 +68,7 @@ def test_hosted_vllm_responses_create_with_string_input():
     Test that hosted_vllm routes directly to the native /v1/responses endpoint
     when the Responses API config is registered, and correctly parses the response.
     """
-    mock_client = _make_mock_http_client(
-        _make_mock_responses_api_response("I'm doing well, thanks!")
-    )
+    mock_client = _make_mock_http_client(_make_mock_responses_api_response("I'm doing well, thanks!"))
 
     with patch(
         "litellm.llms.custom_httpx.llm_http_handler._get_httpx_client",
@@ -109,10 +107,7 @@ def test_hosted_vllm_responses_create_with_explicit_none_extra_body():
     )
 
     # extra_body=None should be normalized to an empty dict (or absent)
-    assert (
-        optional_params.get("extra_body") is not None
-        or "extra_body" not in optional_params
-    )
+    assert optional_params.get("extra_body") is not None or "extra_body" not in optional_params
 
 
 def test_hosted_vllm_provider_config_registration():
diff --git a/tests/test_litellm/llms/hosted_vllm/test_hosted_vllm_rerank_transformation.py b/tests/unit/llms/hosted_vllm/test_hosted_vllm_rerank_transformation.py
similarity index 100%
rename from tests/test_litellm/llms/hosted_vllm/test_hosted_vllm_rerank_transformation.py
rename to tests/unit/llms/hosted_vllm/test_hosted_vllm_rerank_transformation.py
diff --git a/tests/test_litellm/llms/hosted_vllm/videos/test_hosted_vllm_video_transformation.py b/tests/unit/llms/hosted_vllm/videos/test_hosted_vllm_video_transformation.py
similarity index 100%
rename from tests/test_litellm/llms/hosted_vllm/videos/test_hosted_vllm_video_transformation.py
rename to tests/unit/llms/hosted_vllm/videos/test_hosted_vllm_video_transformation.py
diff --git a/tests/test_litellm/llms/huggingface/rerank/test_huggingface_rerank_transformation.py b/tests/unit/llms/huggingface/rerank/test_huggingface_rerank_transformation.py
similarity index 91%
rename from tests/test_litellm/llms/huggingface/rerank/test_huggingface_rerank_transformation.py
rename to tests/unit/llms/huggingface/rerank/test_huggingface_rerank_transformation.py
index 9d6b7290eb6..6fd2b006fef 100644
--- a/tests/test_litellm/llms/huggingface/rerank/test_huggingface_rerank_transformation.py
+++ b/tests/unit/llms/huggingface/rerank/test_huggingface_rerank_transformation.py
@@ -219,29 +219,6 @@ def test_huggingface_rerank_return_documents(mock_post):
             assert "text" in result["document"]
 
 
-@patch("litellm.llms.custom_httpx.http_handler.HTTPHandler.post")
-def test_huggingface_rerank_error_handling(mock_post):
-    """Test HuggingFace rerank error handling."""
-
-    def return_val():
-        return {"error": "Unauthorized"}
-
-    mock_response = MagicMock()
-    mock_response.status_code = 401
-    mock_response.json = return_val
-    mock_response.text = "Unauthorized"
-    mock_post.return_value = mock_response
-
-    with pytest.raises(litellm.APIConnectionError):
-        litellm.rerank(
-            model="huggingface/BAAI/bge-reranker-base",
-            query="hello",
-            documents=["hello", "world"],
-            top_n=2,
-            api_key="invalid_key",
-        )
-
-
 def test_huggingface_rerank_config():
     """Test HuggingFaceRerankConfig class functionality."""
     from litellm.llms.huggingface.rerank.transformation import HuggingFaceRerankConfig
@@ -249,10 +226,7 @@ def test_huggingface_rerank_config():
     config = HuggingFaceRerankConfig()
 
     # Test complete URL generation
-    assert (
-        config.get_complete_url(None, "test")
-        == "https://api-inference.huggingface.co/rerank"
-    )
+    assert config.get_complete_url(None, "test") == "https://api-inference.huggingface.co/rerank"
 
     # Test custom API base
     custom_url = config.get_complete_url("https://custom.huggingface.co", "test")
@@ -292,13 +266,9 @@ def test_request_transformation():
 
     config = HuggingFaceRerankConfig()
 
-    optional_params = OptionalRerankParams(
-        query="hello", texts=["hello", "world"], top_n=2, return_text=True
-    )
+    optional_params = OptionalRerankParams(query="hello", texts=["hello", "world"], top_n=2, return_text=True)
 
-    request_body = config.transform_rerank_request(
-        model="test", optional_rerank_params=optional_params, headers={}
-    )
+    request_body = config.transform_rerank_request(model="test", optional_rerank_params=optional_params, headers={})
 
     assert request_body["query"] == "hello"
     assert request_body["texts"] == ["hello", "world"]
@@ -368,9 +338,7 @@ def test_validate_environment():
 
     # Test headers override
     custom_headers = {"custom": "header"}
-    headers = config.validate_environment(
-        headers=custom_headers, model="test", api_key="test_key"
-    )
+    headers = config.validate_environment(headers=custom_headers, model="test", api_key="test_key")
 
     assert "custom" in headers
     assert headers["custom"] == "header"
diff --git a/tests/test_litellm/llms/inception/test_inception_chat_transformation.py b/tests/unit/llms/inception/test_inception_chat_transformation.py
similarity index 96%
rename from tests/test_litellm/llms/inception/test_inception_chat_transformation.py
rename to tests/unit/llms/inception/test_inception_chat_transformation.py
index 1d12be2adee..c4c023077fc 100644
--- a/tests/test_litellm/llms/inception/test_inception_chat_transformation.py
+++ b/tests/unit/llms/inception/test_inception_chat_transformation.py
@@ -188,21 +188,15 @@ def test_inception_does_not_leak_key_to_caller_api_base():
     caller also supplies their own key.
     """
     config = InceptionChatConfig()
-    with mock.patch.dict(
-        os.environ, {"INCEPTION_API_KEY": "server-secret"}, clear=True
-    ):
+    with mock.patch.dict(os.environ, {"INCEPTION_API_KEY": "server-secret"}, clear=True):
         with mock.patch.object(litellm, "inception_key", "module-secret"):
             # caller overrides api_base without a key -> server key withheld
-            api_base, api_key = config._get_openai_compatible_provider_info(
-                "https://attacker.example/v1", None
-            )
+            api_base, api_key = config._get_openai_compatible_provider_info("https://attacker.example/v1", None)
             assert api_base == "https://attacker.example/v1"
             assert api_key is None
 
             # caller overrides api_base AND supplies their own key -> used as-is
-            _, api_key = config._get_openai_compatible_provider_info(
-                "https://attacker.example/v1", "caller-key"
-            )
+            _, api_key = config._get_openai_compatible_provider_info("https://attacker.example/v1", "caller-key")
             assert api_key == "caller-key"
 
             # default/server base -> server-managed key resolved
@@ -217,9 +211,7 @@ def test_get_llm_provider_inception():
     assert model == "mercury-2"
     assert provider == "inception"
 
-    model, provider, _, api_base = get_llm_provider(
-        "mercury-2", api_base="https://api.inceptionlabs.ai/v1"
-    )
+    model, provider, _, api_base = get_llm_provider("mercury-2", api_base="https://api.inceptionlabs.ai/v1")
     assert model == "mercury-2"
     assert provider == "inception"
     assert api_base == "https://api.inceptionlabs.ai/v1"
@@ -293,5 +285,3 @@ def test_inception_completion_targets_inception_endpoint():
     assert captured["body"]["model"] == "mercury-2"
     assert captured["body"]["tool_choice"] == "auto"
     assert response.choices[0].message.content == "hi"
-
-
diff --git a/tests/test_litellm/llms/inception/test_inception_completion_transformation.py b/tests/unit/llms/inception/test_inception_completion_transformation.py
similarity index 95%
rename from tests/test_litellm/llms/inception/test_inception_completion_transformation.py
rename to tests/unit/llms/inception/test_inception_completion_transformation.py
index ed3f34fc744..84923229e20 100644
--- a/tests/test_litellm/llms/inception/test_inception_completion_transformation.py
+++ b/tests/unit/llms/inception/test_inception_completion_transformation.py
@@ -22,9 +22,7 @@ def _fim_response_bytes():
             "object": "text_completion",
             "created": 1,
             "model": "mercury-edit-2",
-            "choices": [
-                {"text": "a + b", "index": 0, "finish_reason": "stop", "logprobs": None}
-            ],
+            "choices": [{"text": "a + b", "index": 0, "finish_reason": "stop", "logprobs": None}],
             "usage": {"prompt_tokens": 5, "completion_tokens": 3, "total_tokens": 8},
         }
     ).encode()
@@ -47,9 +45,7 @@ def test_inception_fim_supports_suffix_param():
 
 def test_inception_fim_supported_params_match_schema():
     """FIM exposes the OpenAI subset of Inception's FIMCompletionRequest only"""
-    params = InceptionTextCompletionConfig().get_supported_openai_params(
-        "mercury-edit-2"
-    )
+    params = InceptionTextCompletionConfig().get_supported_openai_params("mercury-edit-2")
     for p in ("suffix", "top_p", "frequency_penalty", "presence_penalty", "stop"):
         assert p in params
     # Chat-only sampling controls are not part of Inception's FIM schema
@@ -75,11 +71,7 @@ def test_inception_get_supported_openai_params_dispatch():
 
 @pytest.mark.parametrize("provider", ["inception", "text-completion-inception"])
 def test_inception_validate_environment(provider):
-    model = (
-        "inception/mercury-2"
-        if provider == "inception"
-        else "text-completion-inception/mercury-edit-2"
-    )
+    model = "inception/mercury-2" if provider == "inception" else "text-completion-inception/mercury-edit-2"
 
     with mock.patch.dict(os.environ, {}, clear=True):
         result = litellm.validate_environment(model)
@@ -217,9 +209,7 @@ def test_inception_fim_does_not_leak_global_api_key():
             content=_fim_response_bytes(),
         )
 
-    with mock.patch.dict(
-        os.environ, {"INCEPTION_API_KEY": "sk-inception-correct"}, clear=True
-    ):
+    with mock.patch.dict(os.environ, {"INCEPTION_API_KEY": "sk-inception-correct"}, clear=True):
         with mock.patch.object(litellm, "inception_key", None):
             with mock.patch.object(litellm, "api_key", "sk-global-should-not-leak"):
                 with mock.patch("httpx.Client.send", new=fake_send):
diff --git a/tests/test_litellm/llms/jina_ai/embedding/test_jina_embedding_transformation.py b/tests/unit/llms/jina_ai/embedding/test_jina_embedding_transformation.py
similarity index 100%
rename from tests/test_litellm/llms/jina_ai/embedding/test_jina_embedding_transformation.py
rename to tests/unit/llms/jina_ai/embedding/test_jina_embedding_transformation.py
diff --git a/tests/test_litellm/llms/langflow/chat/test_langflow_chat_transformation.py b/tests/unit/llms/langflow/chat/test_langflow_chat_transformation.py
similarity index 93%
rename from tests/test_litellm/llms/langflow/chat/test_langflow_chat_transformation.py
rename to tests/unit/llms/langflow/chat/test_langflow_chat_transformation.py
index 383a7afbe93..179a6cad4aa 100644
--- a/tests/test_litellm/llms/langflow/chat/test_langflow_chat_transformation.py
+++ b/tests/unit/llms/langflow/chat/test_langflow_chat_transformation.py
@@ -46,7 +46,7 @@ def test_langflow_config_get_complete_url():
 
 def test_langflow_config_get_complete_url_requires_api_base():
     config = LangFlowConfig()
-    with pytest.raises(ValueError, match='api_base is required for LangFlow\\. Set it via'):
+    with pytest.raises(ValueError, match="api_base is required for LangFlow\\. Set it via"):
         config.get_complete_url(
             api_base=None,
             api_key=None,
@@ -225,9 +225,7 @@ def test_langflow_extra_body_cannot_inject_tweaks_into_run_payload():
         posted_bodies.append(json.loads(body) if isinstance(body, str) else body)
         resp = MagicMock(spec=httpx.Response)
         resp.status_code = 200
-        resp.json.return_value = {
-            "outputs": [{"outputs": [{"results": {"message": {"text": "hi"}}}]}]
-        }
+        resp.json.return_value = {"outputs": [{"outputs": [{"results": {"message": {"text": "hi"}}}]}]}
         resp.headers = {}
         resp.text = "{}"
         return resp
@@ -275,9 +273,7 @@ def test_langflow_config_extract_response_from_outputs_dict():
                     "outputs": [
                         {
                             "results": {},
-                            "outputs": {
-                                "message": {"message": {"text": "via outputs dict"}}
-                            },
+                            "outputs": {"message": {"message": {"text": "via outputs dict"}}},
                         }
                     ]
                 }
@@ -292,14 +288,9 @@ def test_langflow_extract_response_returns_none_when_no_message():
     assert config._extract_content_from_response({"outputs": []}) is None
     assert config._extract_content_from_response({"detail": "flow failed"}) is None
     assert config._extract_content_from_response({"outputs": ["not-a-dict"]}) is None
+    assert config._extract_content_from_response({"outputs": [{"outputs": ["bad"]}]}) is None
     assert (
-        config._extract_content_from_response({"outputs": [{"outputs": ["bad"]}]})
-        is None
-    )
-    assert (
-        config._extract_content_from_response(
-            {"outputs": [{"outputs": [{"results": {"message": {"text": ""}}}]}]}
-        )
+        config._extract_content_from_response({"outputs": [{"outputs": [{"results": {"message": {"text": ""}}}]}]})
         is None
     )
 
@@ -310,9 +301,7 @@ def test_langflow_transform_response_builds_model_response_with_usage():
         status_code=200,
         json={
             "session_id": "sess-abc",
-            "outputs": [
-                {"outputs": [{"results": {"message": {"text": "Hello from LangFlow"}}}]}
-            ],
+            "outputs": [{"outputs": [{"results": {"message": {"text": "Hello from LangFlow"}}}]}],
         },
     )
 
@@ -332,9 +321,7 @@ def test_langflow_transform_response_builds_model_response_with_usage():
     assert result.choices[0].finish_reason == "stop"
     assert result.model == "langflow/my-flow-id"
     assert result.usage.completion_tokens > 0
-    assert result.usage.total_tokens == (
-        result.usage.prompt_tokens + result.usage.completion_tokens
-    )
+    assert result.usage.total_tokens == (result.usage.prompt_tokens + result.usage.completion_tokens)
 
 
 def test_langflow_transform_response_raises_on_unparseable_body():
@@ -357,9 +344,7 @@ def test_langflow_transform_response_raises_on_unparseable_body():
 
 def test_langflow_transform_response_raises_on_non_json_body():
     config = LangFlowConfig()
-    raw_response = httpx.Response(
-        status_code=200, content=b"not json", headers={"content-type": "text/plain"}
-    )
+    raw_response = httpx.Response(status_code=200, content=b"not json", headers={"content-type": "text/plain"})
 
     with pytest.raises(LangFlowError):
         config.transform_response(
diff --git a/tests/test_litellm/llms/litellm_proxy/test_sandbox_executor.py b/tests/unit/llms/litellm_proxy/test_sandbox_executor.py
similarity index 84%
rename from tests/test_litellm/llms/litellm_proxy/test_sandbox_executor.py
rename to tests/unit/llms/litellm_proxy/test_sandbox_executor.py
index 422e7a3cf4d..e7a03b9231a 100644
--- a/tests/test_litellm/llms/litellm_proxy/test_sandbox_executor.py
+++ b/tests/unit/llms/litellm_proxy/test_sandbox_executor.py
@@ -55,9 +55,7 @@ def _install_fake_sandbox(monkeypatch, session_cls=_FakeSandboxSession):
 def test_execute_installs_inline_requirements_file(monkeypatch):
     _install_fake_sandbox(monkeypatch)
     executor = SkillsSandboxExecutor()
-    monkeypatch.setattr(
-        executor, "_collect_generated_files", lambda *args, **kwargs: []
-    )
+    monkeypatch.setattr(executor, "_collect_generated_files", lambda *args, **kwargs: [])
 
     requirements = "git+https://example.com/repo.git#egg=foo\n-r extra.txt\n-e ./pkg\n"
     result = executor.execute(
@@ -69,22 +67,15 @@ def test_execute_installs_inline_requirements_file(monkeypatch):
     assert result["success"] is True
 
     created_session = _FakeSandboxSession.last_instance
-    assert created_session.copied_contents[
-        "/sandbox/.litellm_requirements.txt"
-    ] == requirements.encode("utf-8")
-    assert (
-        "pip', 'install', '-r', '.litellm_requirements.txt'"
-        in created_session.run_calls[0]
-    )
+    assert created_session.copied_contents["/sandbox/.litellm_requirements.txt"] == requirements.encode("utf-8")
+    assert "pip', 'install', '-r', '.litellm_requirements.txt'" in created_session.run_calls[0]
     assert "os.chdir('/sandbox')" in created_session.run_calls[1]
 
 
 def test_execute_uses_skill_requirements_txt(monkeypatch):
     _install_fake_sandbox(monkeypatch)
     executor = SkillsSandboxExecutor()
-    monkeypatch.setattr(
-        executor, "_collect_generated_files", lambda *args, **kwargs: []
-    )
+    monkeypatch.setattr(executor, "_collect_generated_files", lambda *args, **kwargs: [])
 
     result = executor.execute(
         code="print('hello')",
@@ -97,9 +88,7 @@ def test_execute_uses_skill_requirements_txt(monkeypatch):
     assert result["success"] is True
 
     created_session = _FakeSandboxSession.last_instance
-    copied_paths = {
-        sandbox_path for _, sandbox_path in created_session.copy_to_runtime_calls
-    }
+    copied_paths = {sandbox_path for _, sandbox_path in created_session.copy_to_runtime_calls}
     assert "/sandbox/requirements.txt" in copied_paths
     assert "/sandbox/.litellm_requirements.txt" not in copied_paths
     assert "pip', 'install', '-r', 'requirements.txt'" in created_session.run_calls[0]
@@ -118,9 +107,7 @@ def test_execute_returns_install_failure(monkeypatch):
 
     _install_fake_sandbox(monkeypatch, session_cls=_FailingSandboxSession)
     executor = SkillsSandboxExecutor()
-    monkeypatch.setattr(
-        executor, "_collect_generated_files", lambda *args, **kwargs: []
-    )
+    monkeypatch.setattr(executor, "_collect_generated_files", lambda *args, **kwargs: [])
 
     result = executor.execute(
         code="print('hello')",
diff --git a/tests/test_litellm/llms/litellm_proxy/test_skills_ownership.py b/tests/unit/llms/litellm_proxy/test_skills_ownership.py
similarity index 88%
rename from tests/test_litellm/llms/litellm_proxy/test_skills_ownership.py
rename to tests/unit/llms/litellm_proxy/test_skills_ownership.py
index e538c50cde8..6caa2da3169 100644
--- a/tests/test_litellm/llms/litellm_proxy/test_skills_ownership.py
+++ b/tests/unit/llms/litellm_proxy/test_skills_ownership.py
@@ -37,12 +37,7 @@ def _skill(skill_id: str, created_by: str | None) -> LiteLLM_SkillsTable:
 def test_should_extract_skill_auth_from_supported_metadata_fields():
     auth = UserAPIKeyAuth(user_id="user-1")
 
-    assert (
-        skills_main._get_user_api_key_auth_from_kwargs(
-            {"metadata": {"user_api_key_auth": auth}}
-        )
-        is auth
-    )
+    assert skills_main._get_user_api_key_auth_from_kwargs({"metadata": {"user_api_key_auth": auth}}) is auth
     assert (
         skills_main._get_user_api_key_auth_from_kwargs(
             {"metadata": {}, "litellm_metadata": {"user_api_key_auth": auth}}
@@ -122,9 +117,7 @@ def test_should_forward_skill_auth_through_sdk_entrypoints(monkeypatch):
         == "deleted"
     )
 
-    assert handler.create_skill_handler.call_args.kwargs["metadata"] == {
-        "source": "request"
-    }
+    assert handler.create_skill_handler.call_args.kwargs["metadata"] == {"source": "request"}
     assert handler.create_skill_handler.call_args.kwargs["user_api_key_dict"] is auth
     assert handler.list_skills_handler.call_args.kwargs["user_api_key_dict"] is auth
     assert handler.get_skill_handler.call_args.kwargs["user_api_key_dict"] is auth
@@ -149,9 +142,7 @@ def test_should_build_resource_owner_scopes_for_auth_context():
     ]
     assert resource_ownership.get_primary_resource_owner_scope(auth) == "user-1"
     assert resource_ownership.user_can_access_resource_owner("team:team-1", auth)
-    assert resource_ownership.get_resource_owner_scopes(
-        UserAPIKeyAuth(token="token-hash")
-    ) == ["key:token-hash"]
+    assert resource_ownership.get_resource_owner_scopes(UserAPIKeyAuth(token="token-hash")) == ["key:token-hash"]
     # Identity-less callers get an empty scope set — sharing a sentinel
     # would collapse every identity-less caller into the same logical
     # owner, which is a cross-tenant data-access primitive.
@@ -165,9 +156,7 @@ def test_should_allow_admin_and_anonymous_resource_owner_paths():
     assert resource_ownership.is_proxy_admin(admin)
     assert resource_ownership.user_can_access_resource_owner(None, admin)
     assert resource_ownership.user_can_access_resource_owner(None, None)
-    assert not resource_ownership.user_can_access_resource_owner(
-        None, UserAPIKeyAuth(user_id="user-1")
-    )
+    assert not resource_ownership.user_can_access_resource_owner(None, UserAPIKeyAuth(user_id="user-1"))
 
 
 @pytest.mark.asyncio
@@ -218,9 +207,7 @@ async def test_should_forward_skill_auth_through_transformation_handler(monkeypa
 async def test_should_store_team_owner_for_keys_without_user_id(monkeypatch):
     table = AsyncMock()
     table.create.side_effect = lambda data: _skill(data["skill_id"], data["created_by"])
-    prisma_client = type(
-        "Prisma", (), {"db": type("DB", (), {"litellm_skillstable": table})()}
-    )()
+    prisma_client = type("Prisma", (), {"db": type("DB", (), {"litellm_skillstable": table})()})()
     monkeypatch.setattr(
         LiteLLMSkillsHandler,
         "_get_prisma_client",
@@ -242,9 +229,7 @@ async def test_should_store_team_owner_for_keys_without_user_id(monkeypatch):
 async def test_should_store_token_owner_for_keys_without_user_team_or_org(monkeypatch):
     table = AsyncMock()
     table.create.side_effect = lambda data: _skill(data["skill_id"], data["created_by"])
-    prisma_client = type(
-        "Prisma", (), {"db": type("DB", (), {"litellm_skillstable": table})()}
-    )()
+    prisma_client = type("Prisma", (), {"db": type("DB", (), {"litellm_skillstable": table})()})()
     monkeypatch.setattr(
         LiteLLMSkillsHandler,
         "_get_prisma_client",
@@ -268,9 +253,7 @@ async def test_should_reject_skill_create_for_identityless_proxy_auth(monkeypatc
     sentinel as ``created_by`` would let any two such callers see each
     other's skills via the resulting shared owner scope."""
     table = AsyncMock()
-    prisma_client = type(
-        "Prisma", (), {"db": type("DB", (), {"litellm_skillstable": table})()}
-    )()
+    prisma_client = type("Prisma", (), {"db": type("DB", (), {"litellm_skillstable": table})()})()
     monkeypatch.setattr(
         LiteLLMSkillsHandler,
         "_get_prisma_client",
@@ -291,9 +274,7 @@ async def test_should_reject_skill_create_for_identityless_proxy_auth(monkeypatc
 async def test_should_filter_list_skills_to_authenticated_owner_scopes(monkeypatch):
     table = AsyncMock()
     table.find_many.return_value = [_skill("litellm_skill_owner", "user-1")]
-    prisma_client = type(
-        "Prisma", (), {"db": type("DB", (), {"litellm_skillstable": table})()}
-    )()
+    prisma_client = type("Prisma", (), {"db": type("DB", (), {"litellm_skillstable": table})()})()
     monkeypatch.setattr(
         LiteLLMSkillsHandler,
         "_get_prisma_client",
@@ -318,9 +299,7 @@ async def test_should_filter_list_skills_to_authenticated_owner_scopes(monkeypat
 async def test_should_hide_skill_from_different_owner(monkeypatch):
     table = AsyncMock()
     table.find_unique.return_value = _skill("litellm_skill_other", "user-2")
-    prisma_client = type(
-        "Prisma", (), {"db": type("DB", (), {"litellm_skillstable": table})()}
-    )()
+    prisma_client = type("Prisma", (), {"db": type("DB", (), {"litellm_skillstable": table})()})()
     monkeypatch.setattr(
         LiteLLMSkillsHandler,
         "_get_prisma_client",
@@ -340,9 +319,7 @@ async def test_should_hide_skill_from_different_owner(monkeypatch):
 async def test_should_hide_unowned_skill_by_default(monkeypatch):
     table = AsyncMock()
     table.find_unique.return_value = _skill("litellm_skill_unowned", None)
-    prisma_client = type(
-        "Prisma", (), {"db": type("DB", (), {"litellm_skillstable": table})()}
-    )()
+    prisma_client = type("Prisma", (), {"db": type("DB", (), {"litellm_skillstable": table})()})()
     monkeypatch.setattr(
         LiteLLMSkillsHandler,
         "_get_prisma_client",
@@ -364,9 +341,7 @@ async def test_list_skills_excludes_unowned_for_non_admin(monkeypatch):
     with ``created_by IS NULL`` are excluded — admin-only."""
     table = AsyncMock()
     table.find_many.return_value = []
-    prisma_client = type(
-        "Prisma", (), {"db": type("DB", (), {"litellm_skillstable": table})()}
-    )()
+    prisma_client = type("Prisma", (), {"db": type("DB", (), {"litellm_skillstable": table})()})()
     monkeypatch.setattr(
         LiteLLMSkillsHandler,
         "_get_prisma_client",
@@ -422,9 +397,7 @@ async def test_load_skill_uses_cache_after_first_db_hit(monkeypatch):
     fake_skill = Mock(created_by="user-1", skill_id="litellm_skill_a")
     table = AsyncMock()
     table.find_unique = AsyncMock(return_value=fake_skill)
-    prisma_client = type(
-        "Prisma", (), {"db": type("DB", (), {"litellm_skillstable": table})()}
-    )()
+    prisma_client = type("Prisma", (), {"db": type("DB", (), {"litellm_skillstable": table})()})()
     monkeypatch.setattr(
         skills_handler.LiteLLMSkillsHandler,
         "_get_prisma_client",
@@ -432,10 +405,7 @@ async def test_load_skill_uses_cache_after_first_db_hit(monkeypatch):
     )
 
     for _ in range(3):
-        assert (
-            await skills_handler.LiteLLMSkillsHandler._load_skill("litellm_skill_a")
-            is fake_skill
-        )
+        assert await skills_handler.LiteLLMSkillsHandler._load_skill("litellm_skill_a") is fake_skill
     assert table.find_unique.await_count == 1
 
 
@@ -445,9 +415,7 @@ async def test_load_skill_caches_negative_lookups(monkeypatch):
     the DB and the caller still sees ``None``."""
     table = AsyncMock()
     table.find_unique = AsyncMock(return_value=None)
-    prisma_client = type(
-        "Prisma", (), {"db": type("DB", (), {"litellm_skillstable": table})()}
-    )()
+    prisma_client = type("Prisma", (), {"db": type("DB", (), {"litellm_skillstable": table})()})()
     monkeypatch.setattr(
         skills_handler.LiteLLMSkillsHandler,
         "_get_prisma_client",
@@ -466,9 +434,7 @@ async def test_delete_skill_invalidates_cache(monkeypatch):
     table = AsyncMock()
     table.find_unique = AsyncMock(return_value=fake_skill)
     table.delete = AsyncMock()
-    prisma_client = type(
-        "Prisma", (), {"db": type("DB", (), {"litellm_skillstable": table})()}
-    )()
+    prisma_client = type("Prisma", (), {"db": type("DB", (), {"litellm_skillstable": table})()})()
     monkeypatch.setattr(
         skills_handler.LiteLLMSkillsHandler,
         "_get_prisma_client",
@@ -480,12 +446,7 @@ async def test_delete_skill_invalidates_cache(monkeypatch):
     assert skills_handler._SKILL_CACHE.get_cache("litellm_skill_a") is fake_skill
 
     auth = UserAPIKeyAuth(user_id="user-1")
-    await skills_handler.LiteLLMSkillsHandler.delete_skill(
-        "litellm_skill_a", user_api_key_dict=auth
-    )
+    await skills_handler.LiteLLMSkillsHandler.delete_skill("litellm_skill_a", user_api_key_dict=auth)
 
     # Post-delete, the cache holds the negative sentinel — not the stale row.
-    assert (
-        skills_handler._SKILL_CACHE.get_cache("litellm_skill_a")
-        == skills_handler._NEGATIVE_SKILL_SENTINEL
-    )
+    assert skills_handler._SKILL_CACHE.get_cache("litellm_skill_a") == skills_handler._NEGATIVE_SKILL_SENTINEL

From d47008129c175a068b23c9e0afe306463e39974f Mon Sep 17 00:00:00 2001
From: yuneng <yuneng@berri.ai>
Date: Sun, 20 Sep 2026 08:08:12 +0000
Subject: [PATCH 03/13] test(llms): migrate phase 7 provider unit tests to
 tests/unit

Move the wave 1 phase 7 batch (fireworks_ai, gemini, gigachat, github_copilot; 20 files) from tests/test_litellm to tests/unit after judging every test function under a behaviour mutation. Seven wiring or mock-echo tests that stayed green are deleted. The fireworks cost calculator tests get a local model_cost save/restore fixture since the tests/unit tree has no shared conftest for it

Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
---
 .../test_fireworks_ai_chat_transformation.py  |  0
 ...test_fireworks_ai_rerank_transformation.py |  0
 ...t_fireworks_ai_responses_transformation.py | 15 --------
 .../test_fireworks_ai_cache_pricing.py        |  0
 .../test_fireworks_ai_common_utils.py         |  0
 .../test_fireworks_ai_cost_calculator.py      | 22 ++++++++---
 ...mini_audio_transcription_transformation.py |  0
 .../files/test_gemini_files_transformation.py |  0
 .../test_google_genai_guardrail_handler.py    |  0
 .../test_gemini_image_edit_transformation.py  |  0
 .../test_gemini_realtime_transformation.py    |  0
 .../test_gemini_video_transformation.py       |  0
 .../chat/test_gigachat_chat_streaming.py      |  0
 .../chat/test_gigachat_chat_transformation.py | 28 --------------
 .../test_gigachat_embedding_transformation.py | 30 ---------------
 ...est_gigachat_passthrough_transformation.py |  0
 .../llms/gigachat/test_authenticator.py       |  0
 .../llms/gigachat/test_file_handler.py        | 38 -------------------
 .../llms/gigachat/test_utils.py               |  0
 ...github_copilot_embedding_transformation.py |  0
 20 files changed, 16 insertions(+), 117 deletions(-)
 rename tests/{test_litellm => unit}/llms/fireworks_ai/chat/test_fireworks_ai_chat_transformation.py (100%)
 rename tests/{test_litellm => unit}/llms/fireworks_ai/rerank/test_fireworks_ai_rerank_transformation.py (100%)
 rename tests/{test_litellm => unit}/llms/fireworks_ai/responses/test_fireworks_ai_responses_transformation.py (97%)
 rename tests/{test_litellm => unit}/llms/fireworks_ai/test_fireworks_ai_cache_pricing.py (100%)
 rename tests/{test_litellm => unit}/llms/fireworks_ai/test_fireworks_ai_common_utils.py (100%)
 rename tests/{test_litellm => unit}/llms/fireworks_ai/test_fireworks_ai_cost_calculator.py (90%)
 rename tests/{test_litellm => unit}/llms/gemini/audio_transcription/test_gemini_audio_transcription_transformation.py (100%)
 rename tests/{test_litellm => unit}/llms/gemini/files/test_gemini_files_transformation.py (100%)
 rename tests/{test_litellm => unit}/llms/gemini/google_genai/guardrail_translation/test_google_genai_guardrail_handler.py (100%)
 rename tests/{test_litellm => unit}/llms/gemini/image_edit/test_gemini_image_edit_transformation.py (100%)
 rename tests/{test_litellm => unit}/llms/gemini/realtime/test_gemini_realtime_transformation.py (100%)
 rename tests/{test_litellm => unit}/llms/gemini/videos/test_gemini_video_transformation.py (100%)
 rename tests/{test_litellm => unit}/llms/gigachat/chat/test_gigachat_chat_streaming.py (100%)
 rename tests/{test_litellm => unit}/llms/gigachat/chat/test_gigachat_chat_transformation.py (95%)
 rename tests/{test_litellm => unit}/llms/gigachat/embedding/test_gigachat_embedding_transformation.py (91%)
 rename tests/{test_litellm => unit}/llms/gigachat/passthrough/test_gigachat_passthrough_transformation.py (100%)
 rename tests/{test_litellm => unit}/llms/gigachat/test_authenticator.py (100%)
 rename tests/{test_litellm => unit}/llms/gigachat/test_file_handler.py (91%)
 rename tests/{test_litellm => unit}/llms/gigachat/test_utils.py (100%)
 rename tests/{test_litellm => unit}/llms/github_copilot/embedding/test_github_copilot_embedding_transformation.py (100%)

diff --git a/tests/test_litellm/llms/fireworks_ai/chat/test_fireworks_ai_chat_transformation.py b/tests/unit/llms/fireworks_ai/chat/test_fireworks_ai_chat_transformation.py
similarity index 100%
rename from tests/test_litellm/llms/fireworks_ai/chat/test_fireworks_ai_chat_transformation.py
rename to tests/unit/llms/fireworks_ai/chat/test_fireworks_ai_chat_transformation.py
diff --git a/tests/test_litellm/llms/fireworks_ai/rerank/test_fireworks_ai_rerank_transformation.py b/tests/unit/llms/fireworks_ai/rerank/test_fireworks_ai_rerank_transformation.py
similarity index 100%
rename from tests/test_litellm/llms/fireworks_ai/rerank/test_fireworks_ai_rerank_transformation.py
rename to tests/unit/llms/fireworks_ai/rerank/test_fireworks_ai_rerank_transformation.py
diff --git a/tests/test_litellm/llms/fireworks_ai/responses/test_fireworks_ai_responses_transformation.py b/tests/unit/llms/fireworks_ai/responses/test_fireworks_ai_responses_transformation.py
similarity index 97%
rename from tests/test_litellm/llms/fireworks_ai/responses/test_fireworks_ai_responses_transformation.py
rename to tests/unit/llms/fireworks_ai/responses/test_fireworks_ai_responses_transformation.py
index d0697ca9b0e..05e3812152e 100644
--- a/tests/test_litellm/llms/fireworks_ai/responses/test_fireworks_ai_responses_transformation.py
+++ b/tests/unit/llms/fireworks_ai/responses/test_fireworks_ai_responses_transformation.py
@@ -406,21 +406,6 @@ def test_responses_call_sends_session_affinity_for_caller_session_id() -> None:
     assert headers["x-session-affinity"] == "sess-42"
 
 
-def test_responses_call_keeps_caller_supplied_session_affinity_header() -> None:
-    client: Final = _mock_http_client(_fireworks_response("accounts/fireworks/models/kimi-k3"))
-    pinned: Final[Mapping[str, str]] = MappingProxyType({"x-session-affinity": "explicit-node"})
-    with patch(HTTPX_CLIENT_FACTORY, return_value=client):
-        litellm.responses(
-            model="fireworks_ai/kimi-k3",
-            input="hi",
-            api_key="fw-test-key",
-            litellm_session_id="sess-42",
-            extra_headers=pinned,
-        )
-    _, headers, _ = _sent_request(client)
-    assert headers["x-session-affinity"] == "explicit-node"
-
-
 def test_responses_call_maps_provider_errors_to_fireworks_ai() -> None:
     client: Final = MagicMock()
     request: Final = httpx.Request("POST", FIREWORKS_RESPONSES_URL)
diff --git a/tests/test_litellm/llms/fireworks_ai/test_fireworks_ai_cache_pricing.py b/tests/unit/llms/fireworks_ai/test_fireworks_ai_cache_pricing.py
similarity index 100%
rename from tests/test_litellm/llms/fireworks_ai/test_fireworks_ai_cache_pricing.py
rename to tests/unit/llms/fireworks_ai/test_fireworks_ai_cache_pricing.py
diff --git a/tests/test_litellm/llms/fireworks_ai/test_fireworks_ai_common_utils.py b/tests/unit/llms/fireworks_ai/test_fireworks_ai_common_utils.py
similarity index 100%
rename from tests/test_litellm/llms/fireworks_ai/test_fireworks_ai_common_utils.py
rename to tests/unit/llms/fireworks_ai/test_fireworks_ai_common_utils.py
diff --git a/tests/test_litellm/llms/fireworks_ai/test_fireworks_ai_cost_calculator.py b/tests/unit/llms/fireworks_ai/test_fireworks_ai_cost_calculator.py
similarity index 90%
rename from tests/test_litellm/llms/fireworks_ai/test_fireworks_ai_cost_calculator.py
rename to tests/unit/llms/fireworks_ai/test_fireworks_ai_cost_calculator.py
index 52222f22a51..c6096ba2745 100644
--- a/tests/test_litellm/llms/fireworks_ai/test_fireworks_ai_cost_calculator.py
+++ b/tests/unit/llms/fireworks_ai/test_fireworks_ai_cost_calculator.py
@@ -1,4 +1,5 @@
 import math
+from collections.abc import Generator
 from datetime import datetime, timezone
 from typing import Final
 
@@ -24,6 +25,15 @@ CACHE_READ_COST = litellm.get_model_info(model=MODEL, custom_llm_provider="firew
 OUTPUT_COST = 4.4e-06
 
 
+@pytest.fixture(autouse=True)
+def restore_model_cost() -> Generator[None, None, None]:
+    original: Final = litellm.model_cost
+    litellm.get_model_info.cache_clear()
+    yield
+    litellm.model_cost = original
+    litellm.get_model_info.cache_clear()
+
+
 def _usage(prompt_tokens: int, cached_tokens: int, completion_tokens: int) -> Usage:
     return Usage(
         prompt_tokens=prompt_tokens,
@@ -57,7 +67,7 @@ def _register_off_peak_model(
     cache_read_cost: float | None = STANDARD_CACHE_READ_COST,
     model: str = OFF_PEAK_MODEL,
 ) -> None:
-    litellm.model_cost = {  # test-quality-ok: conftest restores litellm.model_cost after each test
+    litellm.model_cost = {  # test-quality-ok: the restore_model_cost fixture returns litellm.model_cost to the original object after each test
         **litellm.model_cost,
         f"fireworks_ai/{model}": {
             "litellm_provider": "fireworks_ai",
@@ -151,7 +161,7 @@ def test_an_entry_without_a_cache_read_rate_bills_cached_tokens_at_the_documente
     """Fireworks documents a default 50% cached-token discount for serverless models:
     https://docs.fireworks.ai/guides/prompt-caching, accessed 2026-09-19."""
     model = "accounts/fireworks/models/default-cache-read-test"
-    litellm.model_cost = {  # test-quality-ok: conftest restores litellm.model_cost after each test
+    litellm.model_cost = {  # test-quality-ok: the restore_model_cost fixture returns litellm.model_cost to the original object after each test
         **litellm.model_cost,
         f"fireworks_ai/{model}": {
             "litellm_provider": "fireworks_ai",
@@ -171,7 +181,7 @@ def test_an_entry_without_a_cache_read_rate_bills_cached_tokens_at_the_documente
 
 def test_fireworks_cache_read_rates_match_breakdown_and_caching_savings():
     model = "accounts/fireworks/models/breakdown-cache-read-test"
-    litellm.model_cost = {  # test-quality-ok: the save/restore conftest returns litellm.model_cost to the original object after each test, so replacing the map for this entry leaks nothing
+    litellm.model_cost = {  # test-quality-ok: the restore_model_cost fixture returns litellm.model_cost to the original object after each test, so replacing the map for this entry leaks nothing
         **litellm.model_cost,
         f"fireworks_ai/{model}": {
             "litellm_provider": "fireworks_ai",
@@ -204,7 +214,7 @@ def test_fireworks_cache_read_rates_match_breakdown_and_caching_savings():
 
 def test_generic_cost_per_token_applies_fireworks_cache_read_default_with_or_without_model_info():
     model = "accounts/fireworks/models/generic-cache-read-test"
-    litellm.model_cost = {  # test-quality-ok: conftest restores litellm.model_cost after each test
+    litellm.model_cost = {  # test-quality-ok: the restore_model_cost fixture returns litellm.model_cost to the original object after each test
         **litellm.model_cost,
         f"fireworks_ai/{model}": {
             "litellm_provider": "fireworks_ai",
@@ -257,7 +267,7 @@ COMPONENT_AUDIO_OUT_COST = 6e-06
 
 
 def test_cache_write_reasoning_and_audio_tokens_are_billed_at_their_component_rates():
-    litellm.model_cost = {  # test-quality-ok: the save/restore conftest returns litellm.model_cost to the original object after each test, so replacing the map for this entry leaks nothing
+    litellm.model_cost = {  # test-quality-ok: the restore_model_cost fixture returns litellm.model_cost to the original object after each test, so replacing the map for this entry leaks nothing
         **litellm.model_cost,
         f"fireworks_ai/{COMPONENT_MODEL}": {
             "litellm_provider": "fireworks_ai",
@@ -302,7 +312,7 @@ def test_cache_write_reasoning_and_audio_tokens_are_billed_at_their_component_ra
 
 
 def test_an_entry_without_an_input_rate_gets_no_cache_read_fallback():
-    litellm.model_cost = {  # test-quality-ok: the save/restore conftest returns litellm.model_cost to the original object after each test, so replacing the map for this entry leaks nothing
+    litellm.model_cost = {  # test-quality-ok: the restore_model_cost fixture returns litellm.model_cost to the original object after each test, so replacing the map for this entry leaks nothing
         **litellm.model_cost,  # pyright: ignore[reportUnknownMemberType]  # the SDK types model_cost as dict[Unknown, Unknown]
         "fireworks_ai/accounts/fireworks/models/no-input-rate-test": {
             "litellm_provider": "fireworks_ai",
diff --git a/tests/test_litellm/llms/gemini/audio_transcription/test_gemini_audio_transcription_transformation.py b/tests/unit/llms/gemini/audio_transcription/test_gemini_audio_transcription_transformation.py
similarity index 100%
rename from tests/test_litellm/llms/gemini/audio_transcription/test_gemini_audio_transcription_transformation.py
rename to tests/unit/llms/gemini/audio_transcription/test_gemini_audio_transcription_transformation.py
diff --git a/tests/test_litellm/llms/gemini/files/test_gemini_files_transformation.py b/tests/unit/llms/gemini/files/test_gemini_files_transformation.py
similarity index 100%
rename from tests/test_litellm/llms/gemini/files/test_gemini_files_transformation.py
rename to tests/unit/llms/gemini/files/test_gemini_files_transformation.py
diff --git a/tests/test_litellm/llms/gemini/google_genai/guardrail_translation/test_google_genai_guardrail_handler.py b/tests/unit/llms/gemini/google_genai/guardrail_translation/test_google_genai_guardrail_handler.py
similarity index 100%
rename from tests/test_litellm/llms/gemini/google_genai/guardrail_translation/test_google_genai_guardrail_handler.py
rename to tests/unit/llms/gemini/google_genai/guardrail_translation/test_google_genai_guardrail_handler.py
diff --git a/tests/test_litellm/llms/gemini/image_edit/test_gemini_image_edit_transformation.py b/tests/unit/llms/gemini/image_edit/test_gemini_image_edit_transformation.py
similarity index 100%
rename from tests/test_litellm/llms/gemini/image_edit/test_gemini_image_edit_transformation.py
rename to tests/unit/llms/gemini/image_edit/test_gemini_image_edit_transformation.py
diff --git a/tests/test_litellm/llms/gemini/realtime/test_gemini_realtime_transformation.py b/tests/unit/llms/gemini/realtime/test_gemini_realtime_transformation.py
similarity index 100%
rename from tests/test_litellm/llms/gemini/realtime/test_gemini_realtime_transformation.py
rename to tests/unit/llms/gemini/realtime/test_gemini_realtime_transformation.py
diff --git a/tests/test_litellm/llms/gemini/videos/test_gemini_video_transformation.py b/tests/unit/llms/gemini/videos/test_gemini_video_transformation.py
similarity index 100%
rename from tests/test_litellm/llms/gemini/videos/test_gemini_video_transformation.py
rename to tests/unit/llms/gemini/videos/test_gemini_video_transformation.py
diff --git a/tests/test_litellm/llms/gigachat/chat/test_gigachat_chat_streaming.py b/tests/unit/llms/gigachat/chat/test_gigachat_chat_streaming.py
similarity index 100%
rename from tests/test_litellm/llms/gigachat/chat/test_gigachat_chat_streaming.py
rename to tests/unit/llms/gigachat/chat/test_gigachat_chat_streaming.py
diff --git a/tests/test_litellm/llms/gigachat/chat/test_gigachat_chat_transformation.py b/tests/unit/llms/gigachat/chat/test_gigachat_chat_transformation.py
similarity index 95%
rename from tests/test_litellm/llms/gigachat/chat/test_gigachat_chat_transformation.py
rename to tests/unit/llms/gigachat/chat/test_gigachat_chat_transformation.py
index 2f9511e642c..8e84072e549 100644
--- a/tests/test_litellm/llms/gigachat/chat/test_gigachat_chat_transformation.py
+++ b/tests/unit/llms/gigachat/chat/test_gigachat_chat_transformation.py
@@ -141,22 +141,6 @@ class TestValidateEnvironment:
         assert self.config._current_credentials == "my-creds"
         assert self.config._current_api_base == "https://my-api.example.com"
 
-    @patch(f"{TRANSFORM_MODULE}.get_access_token", return_value="token")
-    @patch(f"{TRANSFORM_MODULE}.get_secret_str")
-    def test_falls_back_to_env_for_credentials(  # test-quality-ok: mock-echo of internal wiring
-        self, mock_get_secret, mock_get_token
-    ):
-        mock_get_secret.return_value = "env-creds"
-        self.config.validate_environment(
-            headers={},
-            model="GigaChat",
-            messages=[],
-            optional_params={},
-            litellm_params={},
-            api_key=None,
-            api_base=None,
-        )
-        mock_get_secret.assert_any_call("GIGACHAT_CREDENTIALS")  # test-quality-ok: mock-echo of internal wiring
 
 
 class TestGetSupportedOpenAiParams:
@@ -865,18 +849,6 @@ class TestUploadImage:
     def setup_method(self):
         self.config = GigaChatConfig()
 
-    @patch(f"{TRANSFORM_MODULE}.upload_file_sync", return_value="file-uploaded")
-    def test_upload_image_success(self, mock_upload):
-        self.config._current_credentials = "creds"
-        self.config._current_api_base = "https://api.example.com"
-        result = self.config._upload_image("https://example.com/img.jpg")
-        assert result == "file-uploaded"
-        mock_upload.assert_called_once_with(
-            image_url="https://example.com/img.jpg",
-            credentials="creds",
-            api_base="https://api.example.com",
-        )
-
     @patch(f"{TRANSFORM_MODULE}.upload_file_sync", side_effect=Exception("fail"))
     def test_upload_image_failure_returns_none(self, mock_upload):
         result = self.config._upload_image("https://example.com/img.jpg")
diff --git a/tests/test_litellm/llms/gigachat/embedding/test_gigachat_embedding_transformation.py b/tests/unit/llms/gigachat/embedding/test_gigachat_embedding_transformation.py
similarity index 91%
rename from tests/test_litellm/llms/gigachat/embedding/test_gigachat_embedding_transformation.py
rename to tests/unit/llms/gigachat/embedding/test_gigachat_embedding_transformation.py
index 8537793ea72..01fe66ca4c7 100644
--- a/tests/test_litellm/llms/gigachat/embedding/test_gigachat_embedding_transformation.py
+++ b/tests/unit/llms/gigachat/embedding/test_gigachat_embedding_transformation.py
@@ -37,17 +37,6 @@ def _make_httpx_response(body: dict, status_code: int = 200) -> httpx.Response:
 # ---------------------------------------------------------------------------
 
 
-class TestGetConfig:
-    def setup_method(self):
-        self.config = GigaChatEmbeddingConfig()
-
-    def test_contains_only_abc_impl(self):
-        """get_config returns ABC internal data due to inheritance."""
-        result = self.config.get_config()
-        # The only key should be _abc_impl from ABC base class
-        assert set(result.keys()) == {"_abc_impl"}
-
-
 class TestGetSupportedOpenAiParams:
     def setup_method(self):
         self.config = GigaChatEmbeddingConfig()
@@ -287,25 +276,6 @@ class TestTransformEmbeddingResponse:
         )
         assert result.model == "Embeddings"
 
-    def test_calls_logging_post_call(self):
-        raw = self._make_gigachat_response([
-            {"object": "embedding", "embedding": [0.1], "index": 0},
-        ])
-        model_response = EmbeddingResponse()
-        self.config.transform_embedding_response(
-            model="gigachat/Embeddings",
-            raw_response=raw,
-            model_response=model_response,
-            logging_obj=self.logging_obj,
-            api_key="test-api-key",
-            request_data={"input": ["hello"]},
-            optional_params={},
-            litellm_params={},
-        )
-        self.logging_obj.post_call.assert_called_once()
-        args = self.logging_obj.post_call.call_args.kwargs
-        assert args["api_key"] == "test-api-key"
-        assert args["input"] == ["hello"]
 
 
 class TestValidateEnvironment:
diff --git a/tests/test_litellm/llms/gigachat/passthrough/test_gigachat_passthrough_transformation.py b/tests/unit/llms/gigachat/passthrough/test_gigachat_passthrough_transformation.py
similarity index 100%
rename from tests/test_litellm/llms/gigachat/passthrough/test_gigachat_passthrough_transformation.py
rename to tests/unit/llms/gigachat/passthrough/test_gigachat_passthrough_transformation.py
diff --git a/tests/test_litellm/llms/gigachat/test_authenticator.py b/tests/unit/llms/gigachat/test_authenticator.py
similarity index 100%
rename from tests/test_litellm/llms/gigachat/test_authenticator.py
rename to tests/unit/llms/gigachat/test_authenticator.py
diff --git a/tests/test_litellm/llms/gigachat/test_file_handler.py b/tests/unit/llms/gigachat/test_file_handler.py
similarity index 91%
rename from tests/test_litellm/llms/gigachat/test_file_handler.py
rename to tests/unit/llms/gigachat/test_file_handler.py
index ce9505f11f2..de83b2ddf5f 100644
--- a/tests/test_litellm/llms/gigachat/test_file_handler.py
+++ b/tests/unit/llms/gigachat/test_file_handler.py
@@ -344,25 +344,6 @@ class TestUploadFileSync:
 
         assert result is None
 
-    @patch(f"{FILE_MODULE}.get_api_base", return_value="https://api.example.com")
-    @patch(f"{FILE_MODULE}.get_access_token", return_value="test-token")
-    @patch(f"{FILE_MODULE}._get_httpx_client")
-    def test_uploads_without_optional_args(
-        self, mock_http_handler_cls, mock_get_token, mock_get_api_base
-    ):
-        """Verify that credentials, api_base, and litellm_params are optional."""
-        mock_client = MagicMock()
-        mock_response = MagicMock()
-        mock_response.json.return_value = {"id": "file-no-args"}
-        mock_response.raise_for_status = MagicMock()
-        mock_client.post.return_value = mock_response
-        mock_http_handler_cls.return_value = mock_client
-
-        result = upload_file_sync(image_url=_RED_PNG_DATA_URL)
-
-        assert result == "file-no-args"
-        # Should still have called get_access_token without args
-        mock_get_token.assert_called_once_with(credentials=None, litellm_params=None)
 
 
 # ---------------------------------------------------------------------------
@@ -483,22 +464,3 @@ class TestUploadFileAsync:
         )
 
         assert result is None
-
-    @pytest.mark.asyncio
-    @patch(f"{FILE_MODULE}.get_api_base", return_value="https://api.example.com")
-    @patch(f"{FILE_MODULE}.get_access_token_async", return_value="test-token-async")
-    @patch(f"{FILE_MODULE}.get_async_httpx_client")
-    async def test_uploads_without_optional_args(
-        self, mock_get_client, mock_get_token, mock_get_api_base
-    ):
-        mock_client = MagicMock()
-        mock_response = MagicMock()
-        mock_response.json = MagicMock(return_value={"id": "async-no-args"})
-        mock_response.raise_for_status = MagicMock()
-        mock_client.post = AsyncMock(return_value=mock_response)
-        mock_get_client.return_value = mock_client
-
-        result = await upload_file_async(image_url=_RED_PNG_DATA_URL)
-
-        assert result == "async-no-args"
-        mock_get_token.assert_called_once_with(credentials=None, litellm_params=None)
\ No newline at end of file
diff --git a/tests/test_litellm/llms/gigachat/test_utils.py b/tests/unit/llms/gigachat/test_utils.py
similarity index 100%
rename from tests/test_litellm/llms/gigachat/test_utils.py
rename to tests/unit/llms/gigachat/test_utils.py
diff --git a/tests/test_litellm/llms/github_copilot/embedding/test_github_copilot_embedding_transformation.py b/tests/unit/llms/github_copilot/embedding/test_github_copilot_embedding_transformation.py
similarity index 100%
rename from tests/test_litellm/llms/github_copilot/embedding/test_github_copilot_embedding_transformation.py
rename to tests/unit/llms/github_copilot/embedding/test_github_copilot_embedding_transformation.py

From feca00248cf78fda0da5a1fa5b7f1a164fbe5608 Mon Sep 17 00:00:00 2001
From: yuneng <yuneng@berri.ai>
Date: Sun, 20 Sep 2026 08:11:56 +0000
Subject: [PATCH 04/13] test(bedrock): isolate host AWS config in realtime and
 rerank unit tests

Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
---
 .../bedrock/realtime/test_bedrock_realtime_handler.py    | 9 +++++++++
 .../rerank/test_bedrock_rerank_header_forwarding.py      | 9 +++++++++
 2 files changed, 18 insertions(+)

diff --git a/tests/unit/llms/bedrock/realtime/test_bedrock_realtime_handler.py b/tests/unit/llms/bedrock/realtime/test_bedrock_realtime_handler.py
index a7f0f64ef68..73a78a94e9f 100644
--- a/tests/unit/llms/bedrock/realtime/test_bedrock_realtime_handler.py
+++ b/tests/unit/llms/bedrock/realtime/test_bedrock_realtime_handler.py
@@ -18,6 +18,15 @@ from litellm.llms.bedrock.realtime.handler import BedrockRealtime
 from litellm.llms.bedrock.realtime.transformation import BedrockRealtimeConfig
 
 
+@pytest.fixture(autouse=True)
+def _isolate_host_aws_config(monkeypatch, tmp_path):
+    monkeypatch.setenv("AWS_SHARED_CREDENTIALS_FILE", str(tmp_path / "credentials"))
+    monkeypatch.setenv("AWS_CONFIG_FILE", str(tmp_path / "config"))
+    monkeypatch.setenv("AWS_EC2_METADATA_DISABLED", "true")
+    for env_var in ("AWS_PROFILE", "AWS_DEFAULT_PROFILE", "AWS_BEARER_TOKEN_BEDROCK", "AWS_REGION_NAME", "AWS_DEFAULT_REGION"):
+        monkeypatch.delenv(env_var, raising=False)
+
+
 class FakePayloadPart:
     def __init__(self, bytes_):
         self.bytes_ = bytes_
diff --git a/tests/unit/llms/bedrock/rerank/test_bedrock_rerank_header_forwarding.py b/tests/unit/llms/bedrock/rerank/test_bedrock_rerank_header_forwarding.py
index 2ea61b5e978..aa93ddb21b8 100644
--- a/tests/unit/llms/bedrock/rerank/test_bedrock_rerank_header_forwarding.py
+++ b/tests/unit/llms/bedrock/rerank/test_bedrock_rerank_header_forwarding.py
@@ -15,6 +15,15 @@ from litellm.llms.bedrock.base_aws_llm import Boto3CredentialsInfo
 from litellm.llms.bedrock.rerank.handler import BedrockRerankHandler
 from litellm.llms.custom_httpx.http_handler import AsyncHTTPHandler, HTTPHandler
 
+
+@pytest.fixture(autouse=True)
+def _isolate_host_aws_config(monkeypatch, tmp_path):
+    monkeypatch.setenv("AWS_SHARED_CREDENTIALS_FILE", str(tmp_path / "credentials"))
+    monkeypatch.setenv("AWS_CONFIG_FILE", str(tmp_path / "config"))
+    monkeypatch.setenv("AWS_EC2_METADATA_DISABLED", "true")
+    for env_var in ("AWS_PROFILE", "AWS_DEFAULT_PROFILE", "AWS_BEARER_TOKEN_BEDROCK", "AWS_REGION_NAME", "AWS_DEFAULT_REGION"):
+        monkeypatch.delenv(env_var, raising=False)
+
 # Mock response for Bedrock rerank
 # Format based on Bedrock rerank API response structure
 bedrock_rerank_response = {

From cf2a9b372cffe0e00e44ee252168384c8f5c058d Mon Sep 17 00:00:00 2001
From: yuneng <yuneng@berri.ai>
Date: Sun, 20 Sep 2026 08:21:51 +0000
Subject: [PATCH 05/13] test(gigachat): cover env credential fallback by its
 resulting auth header

Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
---
 .../chat/test_gigachat_chat_transformation.py | 19 +++++++++++++++++++
 1 file changed, 19 insertions(+)

diff --git a/tests/unit/llms/gigachat/chat/test_gigachat_chat_transformation.py b/tests/unit/llms/gigachat/chat/test_gigachat_chat_transformation.py
index 8e84072e549..b1307f56336 100644
--- a/tests/unit/llms/gigachat/chat/test_gigachat_chat_transformation.py
+++ b/tests/unit/llms/gigachat/chat/test_gigachat_chat_transformation.py
@@ -141,6 +141,25 @@ class TestValidateEnvironment:
         assert self.config._current_credentials == "my-creds"
         assert self.config._current_api_base == "https://my-api.example.com"
 
+    @patch(
+        f"{TRANSFORM_MODULE}.get_access_token",
+        side_effect=lambda credentials, litellm_params: f"token-for-{credentials}",
+    )
+    def test_falls_back_to_env_credentials_when_api_key_missing(
+        self, mock_get_token, monkeypatch: pytest.MonkeyPatch
+    ):
+        monkeypatch.setenv("GIGACHAT_CREDENTIALS", "env-creds")
+        result = self.config.validate_environment(
+            headers={},
+            model="GigaChat",
+            messages=[],
+            optional_params={},
+            litellm_params={},
+            api_key=None,
+            api_base=None,
+        )
+        assert result["Authorization"] == "Bearer token-for-env-creds"
+        assert self.config._current_credentials == "env-creds"
 
 
 class TestGetSupportedOpenAiParams:

From ba629f2537a73f74f5d466e8b7d2703902b74014 Mon Sep 17 00:00:00 2001
From: yuneng <yuneng@berri.ai>
Date: Sun, 20 Sep 2026 09:37:17 +0000
Subject: [PATCH 06/13] test(bedrock): wrap long lines flagged by review in
 migrated unit tests

Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
---
 .../image/test_bedrock_image_prepare_request.py     |  9 ++++++---
 .../test_bedrock_passthrough_transformation.py      | 13 ++++++++++---
 .../realtime/test_bedrock_realtime_handler.py       |  8 +++++++-
 .../rerank/test_bedrock_rerank_header_forwarding.py | 11 +++++++++--
 .../test_bedrock_vector_store_transformation.py     |  3 ++-
 5 files changed, 34 insertions(+), 10 deletions(-)

diff --git a/tests/unit/llms/bedrock/image/test_bedrock_image_prepare_request.py b/tests/unit/llms/bedrock/image/test_bedrock_image_prepare_request.py
index 1575ccb5739..b010db3a840 100644
--- a/tests/unit/llms/bedrock/image/test_bedrock_image_prepare_request.py
+++ b/tests/unit/llms/bedrock/image/test_bedrock_image_prepare_request.py
@@ -11,7 +11,8 @@ def test_bedrock_image_prepare_request_with_arn() -> None:
 
     with (
         patch(
-            "litellm.llms.bedrock.image_generation.image_handler.BedrockImageGeneration._get_boto_credentials_from_optional_params"
+            "litellm.llms.bedrock.image_generation.image_handler.BedrockImageGeneration."
+            "_get_boto_credentials_from_optional_params"
         ),
         patch(
             "litellm.llms.bedrock.image_generation.image_handler.BedrockImageGeneration.get_request_headers"
@@ -31,7 +32,8 @@ def test_bedrock_image_prepare_request_with_arn() -> None:
 
     assert (
         request.endpoint_url
-        == "https://bedrock-runtime.test.com/model/arn%3Aaws%3Abedrock%3Aus-east-1%3A123456789012%3Aapplication-inference-profile%2Fabcdefghi123/invoke"
+        == "https://bedrock-runtime.test.com/model/arn%3Aaws%3Abedrock%3Aus-east-1%3A123456789012"
+        "%3Aapplication-inference-profile%2Fabcdefghi123/invoke"
     )
 
 
@@ -41,7 +43,8 @@ def test_bedrock_image_prepare_request_without_arn() -> None:
 
     with (
         patch(
-            "litellm.llms.bedrock.image_generation.image_handler.BedrockImageGeneration._get_boto_credentials_from_optional_params"
+            "litellm.llms.bedrock.image_generation.image_handler.BedrockImageGeneration."
+            "_get_boto_credentials_from_optional_params"
         ),
         patch(
             "litellm.llms.bedrock.image_generation.image_handler.BedrockImageGeneration.get_request_headers"
diff --git a/tests/unit/llms/bedrock/passthrough/test_bedrock_passthrough_transformation.py b/tests/unit/llms/bedrock/passthrough/test_bedrock_passthrough_transformation.py
index 854ef92fa4b..d1d636a15f7 100644
--- a/tests/unit/llms/bedrock/passthrough/test_bedrock_passthrough_transformation.py
+++ b/tests/unit/llms/bedrock/passthrough/test_bedrock_passthrough_transformation.py
@@ -419,7 +419,9 @@ def test_bedrock_passthrough_model_id_arn_encoding():
         ), f"ARN slash should be encoded, but found unencoded version in: {url_str}"
 
         # Verify the complete expected URL structure
-        expected_encoded_model_id = "arn:aws:bedrock:us-east-1:590183661440:application-inference-profile%2Fb943q2qbl3m7"
+        expected_encoded_model_id = (
+            "arn:aws:bedrock:us-east-1:590183661440:application-inference-profile%2Fb943q2qbl3m7"
+        )
         expected_url = f"https://bedrock-runtime.us-east-1.amazonaws.com/model/{expected_encoded_model_id}/converse"
         assert url_str == expected_url, f"Expected {expected_url}, but got: {url_str}"
 
@@ -515,7 +517,10 @@ def test_bedrock_passthrough_model_id_without_arn():
 def _event_frame(event_type: str, payload: dict) -> bytes:
     def header(name: str, value: str) -> bytes:
         name_b, value_b = name.encode(), value.encode()
-        return struct.pack("!B", len(name_b)) + name_b + struct.pack("!B", 7) + struct.pack("!H", len(value_b)) + value_b
+        return (
+            struct.pack("!B", len(name_b)) + name_b
+            + struct.pack("!B", 7) + struct.pack("!H", len(value_b)) + value_b
+        )
 
     payload_b = json.dumps(payload, separators=(",", ":")).encode()
     headers_b = (
@@ -589,7 +594,9 @@ def _feed(collector: PassthroughStreamCollector, stream: bytes, chunk_size: int
 
 def test_converse_stream_collector_keeps_usage_without_retaining_the_stream():
     texts = [f"tok{i} " for i in range(4000)]
-    stream = _event_frame("messageStart", {"role": "assistant"}) + _text_block(0, texts) + _stream_tail("end_turn", 4000)
+    stream = (
+        _event_frame("messageStart", {"role": "assistant"}) + _text_block(0, texts) + _stream_tail("end_turn", 4000)
+    )
     _feed(_converse_stream_collector(), stream)
 
     tracemalloc.start()
diff --git a/tests/unit/llms/bedrock/realtime/test_bedrock_realtime_handler.py b/tests/unit/llms/bedrock/realtime/test_bedrock_realtime_handler.py
index 73a78a94e9f..3aa827beb80 100644
--- a/tests/unit/llms/bedrock/realtime/test_bedrock_realtime_handler.py
+++ b/tests/unit/llms/bedrock/realtime/test_bedrock_realtime_handler.py
@@ -23,7 +23,13 @@ def _isolate_host_aws_config(monkeypatch, tmp_path):
     monkeypatch.setenv("AWS_SHARED_CREDENTIALS_FILE", str(tmp_path / "credentials"))
     monkeypatch.setenv("AWS_CONFIG_FILE", str(tmp_path / "config"))
     monkeypatch.setenv("AWS_EC2_METADATA_DISABLED", "true")
-    for env_var in ("AWS_PROFILE", "AWS_DEFAULT_PROFILE", "AWS_BEARER_TOKEN_BEDROCK", "AWS_REGION_NAME", "AWS_DEFAULT_REGION"):
+    for env_var in (
+        "AWS_PROFILE",
+        "AWS_DEFAULT_PROFILE",
+        "AWS_BEARER_TOKEN_BEDROCK",
+        "AWS_REGION_NAME",
+        "AWS_DEFAULT_REGION",
+    ):
         monkeypatch.delenv(env_var, raising=False)
 
 
diff --git a/tests/unit/llms/bedrock/rerank/test_bedrock_rerank_header_forwarding.py b/tests/unit/llms/bedrock/rerank/test_bedrock_rerank_header_forwarding.py
index aa93ddb21b8..c40830b238f 100644
--- a/tests/unit/llms/bedrock/rerank/test_bedrock_rerank_header_forwarding.py
+++ b/tests/unit/llms/bedrock/rerank/test_bedrock_rerank_header_forwarding.py
@@ -21,7 +21,13 @@ def _isolate_host_aws_config(monkeypatch, tmp_path):
     monkeypatch.setenv("AWS_SHARED_CREDENTIALS_FILE", str(tmp_path / "credentials"))
     monkeypatch.setenv("AWS_CONFIG_FILE", str(tmp_path / "config"))
     monkeypatch.setenv("AWS_EC2_METADATA_DISABLED", "true")
-    for env_var in ("AWS_PROFILE", "AWS_DEFAULT_PROFILE", "AWS_BEARER_TOKEN_BEDROCK", "AWS_REGION_NAME", "AWS_DEFAULT_REGION"):
+    for env_var in (
+        "AWS_PROFILE",
+        "AWS_DEFAULT_PROFILE",
+        "AWS_BEARER_TOKEN_BEDROCK",
+        "AWS_REGION_NAME",
+        "AWS_DEFAULT_REGION",
+    ):
         monkeypatch.delenv(env_var, raising=False)
 
 # Mock response for Bedrock rerank
@@ -39,7 +45,8 @@ bedrock_rerank_response = {
 test_query = "What is the capital of the United States?"
 test_documents = [
     "Carson City is the capital city of the American state of Nevada.",
-    "The Commonwealth of the Northern Mariana Islands is a group of islands in the Pacific Ocean. Its capital is Saipan.",
+    "The Commonwealth of the Northern Mariana Islands is a group of islands in the Pacific Ocean. "
+    "Its capital is Saipan.",
     "Washington, D.C. is the capital of the United States.",
 ]
 
diff --git a/tests/unit/llms/bedrock/vector_stores/test_bedrock_vector_store_transformation.py b/tests/unit/llms/bedrock/vector_stores/test_bedrock_vector_store_transformation.py
index ab5a2531461..b45e70e31d3 100644
--- a/tests/unit/llms/bedrock/vector_stores/test_bedrock_vector_store_transformation.py
+++ b/tests/unit/llms/bedrock/vector_stores/test_bedrock_vector_store_transformation.py
@@ -46,7 +46,8 @@ def test_transform_search_request_encodes_vector_store_id():
 
     assert (
         url
-        == "https://bedrock-agent-runtime.us-west-2.amazonaws.com/knowledgebases/..%2F..%2Fknowledgebases%2Fother%3Fx%3D1%23frag/retrieve"
+        == "https://bedrock-agent-runtime.us-west-2.amazonaws.com/knowledgebases/..%2F..%2Fknowledgebases%2Fother"
+        "%3Fx%3D1%23frag/retrieve"
     )
     assert body["retrievalQuery"].get("text") == "hello"
 

From f61abe00a6bd3eff086e0fc5fc503e4418850d3e Mon Sep 17 00:00:00 2001
From: yuneng <yuneng@berri.ai>
Date: Sun, 20 Sep 2026 09:58:23 +0000
Subject: [PATCH 07/13] test(llms): wrap remaining lines over 120 chars in
 migrated unit tests

Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
---
 .../guardrail_translation/test_handler.py     | 12 ++++--
 ...drock_mantle_passthrough_transformation.py |  8 +++-
 .../chat/test_bytez_chat_transformation.py    | 42 +++++++++++++++----
 tests/unit/llms/chat/test_converse_handler.py |  5 ++-
 4 files changed, 53 insertions(+), 14 deletions(-)

diff --git a/tests/unit/llms/bedrock/passthrough/guardrail_translation/test_handler.py b/tests/unit/llms/bedrock/passthrough/guardrail_translation/test_handler.py
index dee8366ce2d..da7ed635dcb 100644
--- a/tests/unit/llms/bedrock/passthrough/guardrail_translation/test_handler.py
+++ b/tests/unit/llms/bedrock/passthrough/guardrail_translation/test_handler.py
@@ -1072,7 +1072,8 @@ class TestDeAnonymizeConverseStream:
 
     @pytest.mark.asyncio
     async def test_reasoning_text_delta_de_anonymized(self):
-        """Reasoning deltas carry model output; their text must be guardrailed while the reasoning signature is left untouched."""
+        """Reasoning deltas carry model output; their text must be guardrailed while the
+        reasoning signature is left untouched."""
         stream_bytes = (
             _build_event_stream_frame("messageStart", {"role": "assistant"})
             + _build_event_stream_frame(
@@ -1105,7 +1106,8 @@ class TestDeAnonymizeConverseStream:
 
     @pytest.mark.asyncio
     async def test_tool_use_input_delta_de_anonymized(self):
-        """toolUse.input deltas carry model-generated tool arguments and must be guardrailed instead of being forwarded raw."""
+        """toolUse.input deltas carry model-generated tool arguments and must be
+        guardrailed instead of being forwarded raw."""
         stream_bytes = _build_event_stream_frame(
             "contentBlockDelta",
             {"contentBlockIndex": 0, "delta": {"toolUse": {"input": '{"q":"<PERSON_1>"}'}}},
@@ -1154,7 +1156,8 @@ class TestDeAnonymizeConverseStream:
 
     @pytest.mark.asyncio
     async def test_text_and_reasoning_deltas_de_anonymized_independently(self):
-        """Distinct delta kinds must each be guardrailed and written back into their own field without bleeding the de-anonymized text across kinds."""
+        """Distinct delta kinds must each be guardrailed and written back into their own
+        field without bleeding the de-anonymized text across kinds."""
         captured = {}
 
         async def mock_hook(data, user_api_key_dict, response):
@@ -1192,7 +1195,8 @@ class TestDeAnonymizeConverseStream:
 
     @pytest.mark.asyncio
     async def test_reasoning_signature_only_frame_left_unmodified(self):
-        """A reasoning delta carrying only a signature has no guardrailable text; it must be forwarded untouched and the guardrail must not run."""
+        """A reasoning delta carrying only a signature has no guardrailable text; it must
+        be forwarded untouched and the guardrail must not run."""
         stream_bytes = _build_event_stream_frame(
             "contentBlockDelta",
             {"contentBlockIndex": 0, "delta": {"reasoningContent": {"signature": "sig"}}},
diff --git a/tests/unit/llms/bedrock_mantle/passthrough/test_bedrock_mantle_passthrough_transformation.py b/tests/unit/llms/bedrock_mantle/passthrough/test_bedrock_mantle_passthrough_transformation.py
index 090de0a9d3e..27f6c9a9140 100644
--- a/tests/unit/llms/bedrock_mantle/passthrough/test_bedrock_mantle_passthrough_transformation.py
+++ b/tests/unit/llms/bedrock_mantle/passthrough/test_bedrock_mantle_passthrough_transformation.py
@@ -109,7 +109,13 @@ def test_region_falls_back_to_the_mantle_default_without_any_hint(no_ambient_aws
         ({}, {"AWS_BEARER_TOKEN_BEDROCK": "aws-env-key"}, "aws-env-key"),
     ],
 )
-def test_sign_request_uses_the_deployment_bearer_token(no_ambient_aws, monkeypatch, litellm_params, env, expected_bearer):
+def test_sign_request_uses_the_deployment_bearer_token(
+    no_ambient_aws,
+    monkeypatch,
+    litellm_params,
+    env,
+    expected_bearer,
+):
     for name, value in env.items():
         monkeypatch.setenv(name, value)
     headers, body = BedrockMantlePassthroughConfig().sign_request(
diff --git a/tests/unit/llms/bytez/chat/test_bytez_chat_transformation.py b/tests/unit/llms/bytez/chat/test_bytez_chat_transformation.py
index 440304aeac1..157e2e51175 100644
--- a/tests/unit/llms/bytez/chat/test_bytez_chat_transformation.py
+++ b/tests/unit/llms/bytez/chat/test_bytez_chat_transformation.py
@@ -8,6 +8,32 @@ from litellm.llms.bytez.chat.transformation import BytezChatConfig, API_BASE, ve
 TEST_API_KEY = "MOCK_BYTEZ_API_KEY"
 TEST_MODEL_NAME = "google/gemma-3-4b-it"
 TEST_MODEL = f"bytez/{TEST_MODEL_NAME}"
+CAT_IMAGE_URL = (
+    "https://images.squarespace-cdn.com/content/v1/5452d441e4b0c188b51fef1a/1615326541809-TW01PVTOJ4PXQUX"
+    "VRLHI/male-orange-tabby-cat.jpg"
+)
+KAGGLE_AUDIO_URL = (
+    "https://storage.googleapis.com/kagglesdsdata/datasets/1736753/2838478/dataset/dataset/B_ANI01_MC_FN_"
+    "SIM01_101.wav?X-Goog-Algorithm=GOOG4-RSA-SHA256&X-Goog-Credential=databundle-worker-v2%40kaggle-1616"
+    "07.iam.gserviceaccount.com%2F20250711%2Fauto%2Fstorage%2Fgoog4_request&X-Goog-Date=20250711T192905Z&"
+    "X-Goog-Expires=345600&X-Goog-SignedHeaders=host&X-Goog-Signature=812b4bd6fcf9296f8e34f67664d900a81cf"
+    "81a4c8a4f439ce12befc89b4bef07c2645cab20ce5ba8f6b311dffa85aa05b70b4efbe53bced50a43a5e7622ea1ee0d8cc39"
+    "0679cdc6a6aae2c27f75debc1ce2361c595b3c9e1b8c88e2756ffc6b4f290af7f3dfa7232dc69ccc9a2181be756e0d538250"
+    "f9761a8b05ba1ac6c6b5d946f97a16aa14a5609ae62a2c4713c2077fcd34d129dbcdac6bb543ae547507b1a424e4fd09f817"
+    "000943c11507e0a74c514ec212b17427b7fc9e2ce87a250db1258645e4862a4261e3790fd99c9186148ad0653acd2b6a9468"
+    "adbeb94f17b5a685551037fd2cc9fe72fa405a006c0bd42d03be1e4c0dc4023ed3a77171edff3"
+)
+KAGGLE_VIDEO_URL = (
+    "https://storage.googleapis.com/kagglesdsdata/datasets/3957252/6888743/dog1.mp4?X-Goog-Algorithm=GOOG"
+    "4-RSA-SHA256&X-Goog-Credential=databundle-worker-v2%40kaggle-161607.iam.gserviceaccount.com%2F202507"
+    "11%2Fauto%2Fstorage%2Fgoog4_request&X-Goog-Date=20250711T193025Z&X-Goog-Expires=345600&X-Goog-Signed"
+    "Headers=host&X-Goog-Signature=810961d9abcbc2437954fdf19ef216deb65d3977eb354ec10af0d4644627cc6b143a5f"
+    "c6450996bae1787c09d26334de7cd6ff887510a5ac2a6eed3cfcc6673a47686c84c1f2b0bf543009388d83f2cd9551ad5f72"
+    "084513c6a7acd2c718849a4ebe951ccc5631bed014b0d115225c048b9f5de68673a37db24a98ad39cf3d0ba16fb764bf38eb"
+    "90c78c295c21a4ddac08c3c661b65efd511ccb86bacb87a2e2a97a06f53ea1c64d5dcf274001a61bc20867802549601301d9"
+    "99f5a5b2e49fd444b7db860c68e1c67df6e8edd5ad97171eaafb4fa1462453924ea4d78733be411cb6b5c910d4f829cd7189"
+    "c28dc1b22c8ae2a4da844a0d202e9e64bc7fb17947"
+)
 TEST_MESSAGES = [{"role": "user", "content": "Hello"}]
 
 
@@ -148,7 +174,7 @@ class TestBytezChatConfig:
                             "What color is this cat?",
                             {
                                 "type": "image_url",
-                                "url": "https://images.squarespace-cdn.com/content/v1/5452d441e4b0c188b51fef1a/1615326541809-TW01PVTOJ4PXQUXVRLHI/male-orange-tabby-cat.jpg",
+                                "url": CAT_IMAGE_URL,
                             },
                         ],
                     }
@@ -160,7 +186,7 @@ class TestBytezChatConfig:
                             {"type": "text", "text": "What color is this cat?"},
                             {
                                 "type": "image",
-                                "url": "https://images.squarespace-cdn.com/content/v1/5452d441e4b0c188b51fef1a/1615326541809-TW01PVTOJ4PXQUXVRLHI/male-orange-tabby-cat.jpg",
+                                "url": CAT_IMAGE_URL,
                             },
                         ],
                     }
@@ -174,7 +200,7 @@ class TestBytezChatConfig:
                             {"type": "text", "text": "What color is this cat?"},
                             {
                                 "type": "image_url",
-                                "url": "https://images.squarespace-cdn.com/content/v1/5452d441e4b0c188b51fef1a/1615326541809-TW01PVTOJ4PXQUXVRLHI/male-orange-tabby-cat.jpg",
+                                "url": CAT_IMAGE_URL,
                             },
                         ],
                     }
@@ -186,7 +212,7 @@ class TestBytezChatConfig:
                             {"type": "text", "text": "What color is this cat?"},
                             {
                                 "type": "image",
-                                "url": "https://images.squarespace-cdn.com/content/v1/5452d441e4b0c188b51fef1a/1615326541809-TW01PVTOJ4PXQUXVRLHI/male-orange-tabby-cat.jpg",
+                                "url": CAT_IMAGE_URL,
                             },
                         ],
                     }
@@ -200,7 +226,7 @@ class TestBytezChatConfig:
                             {"type": "text", "text": "What kind of cat meow is this?"},
                             {
                                 "type": "input_audio",
-                                "url": "https://storage.googleapis.com/kagglesdsdata/datasets/1736753/2838478/dataset/dataset/B_ANI01_MC_FN_SIM01_101.wav?X-Goog-Algorithm=GOOG4-RSA-SHA256&X-Goog-Credential=databundle-worker-v2%40kaggle-161607.iam.gserviceaccount.com%2F20250711%2Fauto%2Fstorage%2Fgoog4_request&X-Goog-Date=20250711T192905Z&X-Goog-Expires=345600&X-Goog-SignedHeaders=host&X-Goog-Signature=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",
+                                "url": KAGGLE_AUDIO_URL,
                             },
                         ],
                     }
@@ -212,7 +238,7 @@ class TestBytezChatConfig:
                             {"type": "text", "text": "What kind of cat meow is this?"},
                             {
                                 "type": "audio",
-                                "url": "https://storage.googleapis.com/kagglesdsdata/datasets/1736753/2838478/dataset/dataset/B_ANI01_MC_FN_SIM01_101.wav?X-Goog-Algorithm=GOOG4-RSA-SHA256&X-Goog-Credential=databundle-worker-v2%40kaggle-161607.iam.gserviceaccount.com%2F20250711%2Fauto%2Fstorage%2Fgoog4_request&X-Goog-Date=20250711T192905Z&X-Goog-Expires=345600&X-Goog-SignedHeaders=host&X-Goog-Signature=812b4bd6fcf9296f8e34f67664d900a81cf81a4c8a4f439ce12befc89b4bef07c2645cab20ce5ba8f6b311dffa85aa05b70b4efbe53bced50a43a5e7622ea1ee0d8cc390679cdc6a6aae2c27f75debc1ce2361c595b3c9e1b8c88e2756ffc6b4f290af7f3dfa7232dc69ccc9a2181be756e0d538250f9761a8b05ba1ac6c6b5d946f97a16aa14a5609ae62a2c4713c2077fcd34d129dbcdac6bb543ae547507b1a424e4fd09f817000943c11507e0a74c514ec212b17427b7fc9e2ce87a250db1258645e4862a4261e3790fd99c9186148ad0653acd2b6a9468adbeb94f17b5a685551037fd2cc9fe72fa405a006c0bd42d03be1e4c0dc4023ed3a77171edff3",
+                                "url": KAGGLE_AUDIO_URL,
                             },
                         ],
                     }
@@ -226,7 +252,7 @@ class TestBytezChatConfig:
                             {"type": "text", "text": "What kind of dog is this?"},
                             {
                                 "type": "video_url",
-                                "url": "https://storage.googleapis.com/kagglesdsdata/datasets/3957252/6888743/dog1.mp4?X-Goog-Algorithm=GOOG4-RSA-SHA256&X-Goog-Credential=databundle-worker-v2%40kaggle-161607.iam.gserviceaccount.com%2F20250711%2Fauto%2Fstorage%2Fgoog4_request&X-Goog-Date=20250711T193025Z&X-Goog-Expires=345600&X-Goog-SignedHeaders=host&X-Goog-Signature=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",
+                                "url": KAGGLE_VIDEO_URL,
                             },
                         ],
                     }
@@ -238,7 +264,7 @@ class TestBytezChatConfig:
                             {"type": "text", "text": "What kind of dog is this?"},
                             {
                                 "type": "video",
-                                "url": "https://storage.googleapis.com/kagglesdsdata/datasets/3957252/6888743/dog1.mp4?X-Goog-Algorithm=GOOG4-RSA-SHA256&X-Goog-Credential=databundle-worker-v2%40kaggle-161607.iam.gserviceaccount.com%2F20250711%2Fauto%2Fstorage%2Fgoog4_request&X-Goog-Date=20250711T193025Z&X-Goog-Expires=345600&X-Goog-SignedHeaders=host&X-Goog-Signature=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",
+                                "url": KAGGLE_VIDEO_URL,
                             },
                         ],
                     }
diff --git a/tests/unit/llms/chat/test_converse_handler.py b/tests/unit/llms/chat/test_converse_handler.py
index 12b5f03aedc..05debee0602 100644
--- a/tests/unit/llms/chat/test_converse_handler.py
+++ b/tests/unit/llms/chat/test_converse_handler.py
@@ -106,7 +106,10 @@ class TestBedrockRegionInModelPath:
         ), f"modelId mismatch for {model!r}: got {model_id!r}, expected {expected_model_id!r}"
         assert (
             optional_params.get("aws_region_name") == expected_region
-        ), f"region mismatch for {model!r}: got {optional_params.get('aws_region_name')!r}, expected {expected_region!r}"
+        ), (
+            f"region mismatch for {model!r}: "
+            f"got {optional_params.get('aws_region_name')!r}, expected {expected_region!r}"
+        )
 
     def test_explicit_aws_region_name_not_overridden(self):
         """

From ccd8997b0846469ff0e624b2cd837c4d0f8a3da6 Mon Sep 17 00:00:00 2001
From: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Date: Sun, 20 Sep 2026 10:16:48 +0000
Subject: [PATCH 08/13] refactor(types): replace Any with proven types in 34
 files

Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
---
 .../bedrock_agentcore/transformation.py       |  2 +-
 .../providers/watsonx_orchestrate/config.py   |  4 ++--
 litellm/a2a_protocol/utils.py                 |  6 ++---
 .../gitlab/gitlab_prompt_manager.py           |  4 ++--
 litellm/integrations/otel/presets/agentops.py |  7 ++++--
 .../integrations/vantage/vantage_logger.py    |  4 ++--
 litellm/interactions/agents/http_handler.py   | 22 +++++++++----------
 litellm/litellm_core_utils/logging_utils.py   |  4 +++-
 litellm/litellm_core_utils/url_utils.py       |  4 ++--
 litellm/llms/anthropic/chat/handler.py        |  2 +-
 litellm/llms/azure/completion/handler.py      |  9 ++++----
 .../document_intelligence/transformation.py   |  8 +++----
 .../guardrail_translation/base_translation.py |  2 +-
 .../llms/bedrock/batches/transformation.py    | 10 ++++-----
 ...mazon_twelvelabs_pegasus_transformation.py |  2 +-
 litellm/llms/bytez/chat/transformation.py     |  6 ++---
 litellm/llms/custom_httpx/aiohttp_handler.py  |  4 ++--
 .../llms/deprecated_providers/aleph_alpha.py  |  2 +-
 litellm/llms/lemonade/chat/transformation.py  |  2 +-
 .../llms/openai/image_edit/transformation.py  |  2 +-
 .../runwayml/text_to_speech/transformation.py |  2 +-
 .../llms/vertex_ai/files/transformation.py    |  6 ++---
 .../batch_embed_content_handler.py            |  4 ++--
 .../llms/vertex_ai/vertex_ai_non_gemini.py    |  2 +-
 .../audio_transcription/transformation.py     |  4 ++--
 litellm/proxy/a2a/agent_card.py               | 16 +++++++-------
 .../proxy/agent_endpoints/a2a_endpoints.py    |  2 +-
 litellm/proxy/client/credentials.py           |  5 +++--
 .../guardrails_ai/guardrails_ai.py            |  4 ++--
 .../guardrail_hooks/singulr/singulr.py        |  6 ++---
 .../object_permission_utils.py                | 13 ++++++-----
 .../proxy/policy_engine/pipeline_executor.py  | 12 +++++-----
 .../router_strategy/adaptive_router/hooks.py  |  2 +-
 .../auto_router/auto_router.py                |  4 ++--
 34 files changed, 98 insertions(+), 90 deletions(-)

diff --git a/litellm/a2a_protocol/providers/bedrock_agentcore/transformation.py b/litellm/a2a_protocol/providers/bedrock_agentcore/transformation.py
index 9fa9db48af8..c486f1f6d95 100644
--- a/litellm/a2a_protocol/providers/bedrock_agentcore/transformation.py
+++ b/litellm/a2a_protocol/providers/bedrock_agentcore/transformation.py
@@ -85,7 +85,7 @@ def _filter_reserved_headers(
 
 
 def _request_scoped_runtime_session_id(
-    params: Mapping[str, Any],
+    params: Mapping[str, object],
     litellm_params: Mapping[str, Any],
 ) -> str | None:
     context_id: Final = get_session_id_from_a2a_params(params)
diff --git a/litellm/a2a_protocol/providers/watsonx_orchestrate/config.py b/litellm/a2a_protocol/providers/watsonx_orchestrate/config.py
index ca84d3e07b4..44873edf271 100644
--- a/litellm/a2a_protocol/providers/watsonx_orchestrate/config.py
+++ b/litellm/a2a_protocol/providers/watsonx_orchestrate/config.py
@@ -20,7 +20,7 @@ class WatsonxOrchestrateA2AConfig(BaseA2AProviderConfig):
         params: dict[str, Any],
         api_base: str | None = None,
         **kwargs: Any,
-    ) -> dict[str, Any]:
+    ) -> dict[str, object]:
         """Handle a non-streaming A2A request via WXO runs API."""
         litellm_params: Final = kwargs.get("litellm_params")
         if not litellm_params:
@@ -40,7 +40,7 @@ class WatsonxOrchestrateA2AConfig(BaseA2AProviderConfig):
         params: dict[str, Any],
         api_base: str | None = None,
         **kwargs: Any,
-    ) -> AsyncIterator[dict[str, Any]]:
+    ) -> AsyncIterator[dict[str, object]]:
         """Handle a streaming A2A request via WXO streaming runs API."""
         litellm_params: Final = kwargs.get("litellm_params")
         if not litellm_params:
diff --git a/litellm/a2a_protocol/utils.py b/litellm/a2a_protocol/utils.py
index 47f561068cd..7400844bf28 100644
--- a/litellm/a2a_protocol/utils.py
+++ b/litellm/a2a_protocol/utils.py
@@ -17,7 +17,7 @@ class A2ARequestUtils:
     """Utility class for A2A request/response processing."""
 
     @staticmethod
-    def extract_text_from_message(message: Any) -> str:
+    def extract_text_from_message(message: object) -> str:
         """
         Extract text content from A2A message parts.
 
@@ -142,7 +142,7 @@ class A2ARequestUtils:
         return prompt_tokens, completion_tokens, total_tokens
 
 
-def get_session_id_from_a2a_params(params: Mapping[str, Any]) -> str | None:
+def get_session_id_from_a2a_params(params: Mapping[str, object]) -> str | None:
     message: Final = params.get("message", {})
     if isinstance(message, dict):
         return message.get("contextId")
@@ -166,7 +166,7 @@ def scope_session_to_principal(session_id: str, principal: str | None) -> str:
 
 
 # Backwards compatibility aliases
-def extract_text_from_a2a_message(message: Any) -> str:
+def extract_text_from_a2a_message(message: object) -> str:
     return A2ARequestUtils.extract_text_from_message(message)
 
 
diff --git a/litellm/integrations/gitlab/gitlab_prompt_manager.py b/litellm/integrations/gitlab/gitlab_prompt_manager.py
index d4602176650..817d280074f 100644
--- a/litellm/integrations/gitlab/gitlab_prompt_manager.py
+++ b/litellm/integrations/gitlab/gitlab_prompt_manager.py
@@ -200,8 +200,8 @@ class GitLabTemplateManager:
             metadata=metadata,
         )
 
-    def _parse_yaml_basic(self, yaml_str: str) -> dict[str, Any]:
-        result: Final[dict[str, Any]] = {}
+    def _parse_yaml_basic(self, yaml_str: str) -> dict[str, bool | int | float | str]:
+        result: Final[dict[str, bool | int | float | str]] = {}
         for line in yaml_str.split("\n"):
             line = line.strip()
             if ":" in line and not line.startswith("#"):
diff --git a/litellm/integrations/otel/presets/agentops.py b/litellm/integrations/otel/presets/agentops.py
index 965213f2ee4..58123656caa 100644
--- a/litellm/integrations/otel/presets/agentops.py
+++ b/litellm/integrations/otel/presets/agentops.py
@@ -9,9 +9,12 @@ this preset registers a custom exporter (``kind="agentops"``) that mints the JWT
 worker thread, off any event loop — and caches it for the process lifetime.
 """
 
+from collections.abc import Sequence
 from typing import Any, Final
 
 import httpx
+from opentelemetry.sdk.trace import ReadableSpan
+from opentelemetry.sdk.trace.export import SpanExporter, SpanExportResult
 from pydantic import Field
 from pydantic_settings import BaseSettings, SettingsConfigDict
 
@@ -71,7 +74,7 @@ def agentops_preset(
     )
 
 
-def _build_agentops_exporter(spec: ExporterSpec) -> Any:
+def _build_agentops_exporter(spec: ExporterSpec) -> SpanExporter:
     """Factory for the ``agentops`` exporter kind: a lazy-auth OTLP/HTTP exporter."""
     from opentelemetry.exporter.otlp.proto.http.trace_exporter import (
         OTLPSpanExporter,
@@ -106,7 +109,7 @@ def _build_agentops_exporter(spec: ExporterSpec) -> Any:
             except Exception as e:
                 verbose_logger.debug("AgentOps JWT fetch failed: %s", e)
 
-        def export(self, spans: Any) -> Any:
+        def export(self, spans: Sequence[ReadableSpan]) -> SpanExportResult:
             self._ensure_authenticated()
             return super().export(spans)
 
diff --git a/litellm/integrations/vantage/vantage_logger.py b/litellm/integrations/vantage/vantage_logger.py
index c219ba392ab..48a492fdd72 100644
--- a/litellm/integrations/vantage/vantage_logger.py
+++ b/litellm/integrations/vantage/vantage_logger.py
@@ -59,7 +59,7 @@ class VantageLogger(FocusLogger):
                     raw_interval,
                 )
 
-        destination_config: Final[dict[str, Any]] = {}
+        destination_config: Final[dict[str, str]] = {}
         if resolved_api_key:
             destination_config["api_key"] = resolved_api_key
         if resolved_token:
@@ -93,7 +93,7 @@ class VantageLogger(FocusLogger):
 
         pod_lock_manager = None
         if proxy_logging_obj is not None:
-            writer: Final = getattr(proxy_logging_obj, "db_spend_update_writer", None)
+            writer: Final[object] = getattr(proxy_logging_obj, "db_spend_update_writer", None)
             if writer is not None:
                 pod_lock_manager = getattr(writer, "pod_lock_manager", None)
 
diff --git a/litellm/interactions/agents/http_handler.py b/litellm/interactions/agents/http_handler.py
index ec9df0fb488..2afedc34d36 100644
--- a/litellm/interactions/agents/http_handler.py
+++ b/litellm/interactions/agents/http_handler.py
@@ -7,7 +7,7 @@ duplicated. BaseAgentsAPIConfig stays as pure transform code.
 """
 
 from collections.abc import Coroutine, Mapping
-from typing import Any, Final
+from typing import Final
 
 import httpx
 
@@ -38,7 +38,7 @@ class AgentsHTTPHandler(InteractionsHTTPHandler):
         name: str,
         litellm_params: GenericLiteLLMParams,
         logging_obj: LiteLLMLoggingObj,
-        extra_headers: dict[str, Any] | None = None,
+        extra_headers: dict[str, str] | None = None,
         extra_body: Mapping[str, object] | None = None,
         timeout: float | httpx.Timeout | None = None,
         client: HTTPHandler | None = None,
@@ -93,7 +93,7 @@ class AgentsHTTPHandler(InteractionsHTTPHandler):
         name: str,
         litellm_params: GenericLiteLLMParams,
         logging_obj: LiteLLMLoggingObj,
-        extra_headers: dict[str, Any] | None = None,
+        extra_headers: dict[str, str] | None = None,
         extra_body: Mapping[str, object] | None = None,
         timeout: float | httpx.Timeout | None = None,
         client: AsyncHTTPHandler | None = None,
@@ -141,7 +141,7 @@ class AgentsHTTPHandler(InteractionsHTTPHandler):
         agents_api_config: BaseAgentsAPIConfig,
         litellm_params: GenericLiteLLMParams,
         logging_obj: LiteLLMLoggingObj,
-        extra_headers: dict[str, Any] | None = None,
+        extra_headers: dict[str, str] | None = None,
         timeout: float | httpx.Timeout | None = None,
         client: HTTPHandler | None = None,
         _is_async: bool = False,
@@ -181,7 +181,7 @@ class AgentsHTTPHandler(InteractionsHTTPHandler):
         agents_api_config: BaseAgentsAPIConfig,
         litellm_params: GenericLiteLLMParams,
         logging_obj: LiteLLMLoggingObj,
-        extra_headers: dict[str, Any] | None = None,
+        extra_headers: dict[str, str] | None = None,
         timeout: float | httpx.Timeout | None = None,
         client: AsyncHTTPHandler | None = None,
     ) -> AgentListResponse:
@@ -216,7 +216,7 @@ class AgentsHTTPHandler(InteractionsHTTPHandler):
         name: str,
         litellm_params: GenericLiteLLMParams,
         logging_obj: LiteLLMLoggingObj,
-        extra_headers: dict[str, Any] | None = None,
+        extra_headers: dict[str, str] | None = None,
         timeout: float | httpx.Timeout | None = None,
         client: HTTPHandler | None = None,
         _is_async: bool = False,
@@ -259,7 +259,7 @@ class AgentsHTTPHandler(InteractionsHTTPHandler):
         name: str,
         litellm_params: GenericLiteLLMParams,
         logging_obj: LiteLLMLoggingObj,
-        extra_headers: dict[str, Any] | None = None,
+        extra_headers: dict[str, str] | None = None,
         timeout: float | httpx.Timeout | None = None,
         client: AsyncHTTPHandler | None = None,
     ) -> AgentCreateResponse:
@@ -295,7 +295,7 @@ class AgentsHTTPHandler(InteractionsHTTPHandler):
         name: str,
         litellm_params: GenericLiteLLMParams,
         logging_obj: LiteLLMLoggingObj,
-        extra_headers: dict[str, Any] | None = None,
+        extra_headers: dict[str, str] | None = None,
         timeout: float | httpx.Timeout | None = None,
         client: HTTPHandler | None = None,
         _is_async: bool = False,
@@ -338,7 +338,7 @@ class AgentsHTTPHandler(InteractionsHTTPHandler):
         name: str,
         litellm_params: GenericLiteLLMParams,
         logging_obj: LiteLLMLoggingObj,
-        extra_headers: dict[str, Any] | None = None,
+        extra_headers: dict[str, str] | None = None,
         timeout: float | httpx.Timeout | None = None,
         client: AsyncHTTPHandler | None = None,
     ) -> AgentDeleteResult:
@@ -374,7 +374,7 @@ class AgentsHTTPHandler(InteractionsHTTPHandler):
         name: str,
         litellm_params: GenericLiteLLMParams,
         logging_obj: LiteLLMLoggingObj,
-        extra_headers: dict[str, Any] | None = None,
+        extra_headers: dict[str, str] | None = None,
         timeout: float | httpx.Timeout | None = None,
         client: HTTPHandler | None = None,
         _is_async: bool = False,
@@ -417,7 +417,7 @@ class AgentsHTTPHandler(InteractionsHTTPHandler):
         name: str,
         litellm_params: GenericLiteLLMParams,
         logging_obj: LiteLLMLoggingObj,
-        extra_headers: dict[str, Any] | None = None,
+        extra_headers: dict[str, str] | None = None,
         timeout: float | httpx.Timeout | None = None,
         client: AsyncHTTPHandler | None = None,
     ) -> AgentVersionsResponse:
diff --git a/litellm/litellm_core_utils/logging_utils.py b/litellm/litellm_core_utils/logging_utils.py
index 5be9dd7be2f..38a501ecaae 100644
--- a/litellm/litellm_core_utils/logging_utils.py
+++ b/litellm/litellm_core_utils/logging_utils.py
@@ -67,7 +67,9 @@ def _truncate_base64_in_string(value: str) -> str:
     return _DATA_URI_RE.sub(_base64_data_uri_replacer, value)
 
 
-def _truncate_base64_in_value(value: Any) -> Any:
+def _truncate_base64_in_value(
+    value: str | dict[str, object] | list[object] | None,
+) -> str | dict[str, object] | list[object] | None:
     """Iteratively truncate base64 data URIs in a JSON-like value (str/list/dict).
 
     Uses an explicit stack instead of recursion to satisfy the project's
diff --git a/litellm/litellm_core_utils/url_utils.py b/litellm/litellm_core_utils/url_utils.py
index fa070a648f5..b94d5a6886d 100644
--- a/litellm/litellm_core_utils/url_utils.py
+++ b/litellm/litellm_core_utils/url_utils.py
@@ -418,7 +418,7 @@ def _extract_redirect_url(response: httpx.Response, request_url: str) -> str:
     return str(httpx.URL(request_url).join(location))
 
 
-def safe_get(client: Any, url: str, **kwargs: Any) -> httpx.Response:
+def safe_get(client: _UrlFetcher, url: str, **kwargs: Any) -> httpx.Response:
     """
     Fetch a user-supplied URL with SSRF protection on every redirect hop.
 
@@ -461,7 +461,7 @@ def safe_get(client: Any, url: str, **kwargs: Any) -> httpx.Response:
     raise SSRFError("Too many redirects")
 
 
-async def async_safe_get(client: Any, url: str, **kwargs: Any) -> httpx.Response:
+async def async_safe_get(client: _AsyncUrlFetcher, url: str, **kwargs: Any) -> httpx.Response:
     """Async version of safe_get."""
     if not getattr(litellm, "user_url_validation", True):
         kwargs.setdefault("follow_redirects", True)
diff --git a/litellm/llms/anthropic/chat/handler.py b/litellm/llms/anthropic/chat/handler.py
index 359b8bb08c9..ef0f45d8f8b 100644
--- a/litellm/llms/anthropic/chat/handler.py
+++ b/litellm/llms/anthropic/chat/handler.py
@@ -596,7 +596,7 @@ class ModelResponseIterator:
         self.reasoning_content_chunks: list[str] = []
 
         # Track server tool use inputs and results for code_interpreter_results
-        self._server_tool_inputs: dict[str, Any] = {}
+        self._server_tool_inputs: dict[str, object] = {}
         self.tool_results: list[dict[str, Any]] = []
         self._current_server_tool_id: str | None = None
         self._container_id: str | None = None
diff --git a/litellm/llms/azure/completion/handler.py b/litellm/llms/azure/completion/handler.py
index 80934e994f6..23eef51e7ee 100644
--- a/litellm/llms/azure/completion/handler.py
+++ b/litellm/llms/azure/completion/handler.py
@@ -1,6 +1,7 @@
 from collections.abc import Callable
-from typing import Any, Final
+from typing import Final
 
+import httpx
 from openai import AsyncAzureOpenAI, AzureOpenAI
 
 from litellm.litellm_core_utils.litellm_logging import Logging as LiteLLMLoggingObj
@@ -191,7 +192,7 @@ class AzureTextCompletion(BaseAzureLLM):
         model: str,
         api_base: str,
         data: dict,
-        timeout: Any,
+        timeout: float | httpx.Timeout | None,
         model_response: ModelResponse,
         logging_obj: LiteLLMLoggingObj,
         max_retries: int,
@@ -253,7 +254,7 @@ class AzureTextCompletion(BaseAzureLLM):
         api_version: str,
         data: dict,
         model: str,
-        timeout: Any,
+        timeout: float | httpx.Timeout | None,
         azure_ad_token: str | None = None,
         client=None,
         litellm_params: dict = {},
@@ -306,7 +307,7 @@ class AzureTextCompletion(BaseAzureLLM):
         api_version: str,
         data: dict,
         model: str,
-        timeout: Any,
+        timeout: float | httpx.Timeout | None,
         azure_ad_token: str | None = None,
         client=None,
         litellm_params: dict = {},
diff --git a/litellm/llms/azure_ai/ocr/document_intelligence/transformation.py b/litellm/llms/azure_ai/ocr/document_intelligence/transformation.py
index 3a2af8a5aba..23f532be757 100644
--- a/litellm/llms/azure_ai/ocr/document_intelligence/transformation.py
+++ b/litellm/llms/azure_ai/ocr/document_intelligence/transformation.py
@@ -12,7 +12,7 @@ import asyncio
 import re
 import time
 from collections.abc import Mapping
-from typing import TYPE_CHECKING, Any, Final
+from typing import TYPE_CHECKING, Final
 from urllib.parse import quote
 
 import httpx
@@ -127,7 +127,7 @@ class AzureDocumentIntelligenceOCRConfig(BaseOCRConfig):
 
     def map_ocr_params(
         self,
-        non_default_params: dict,
+        non_default_params: Mapping[str, object],
         optional_params: dict,
         model: str,
     ) -> dict:
@@ -164,7 +164,7 @@ class AzureDocumentIntelligenceOCRConfig(BaseOCRConfig):
             raise UnsupportedParamsError(message=f"{e}", model=model, llm_provider="azure_ai") from e
 
     @staticmethod
-    def _normalize_pages_param(pages: Any) -> str:
+    def _normalize_pages_param(pages: object) -> str:
         """
         Convert a caller-provided `pages` value to Azure DI's query-string
         form. Azure expects 1-based page numbers, grammar: `^(\\d+(-\\d+)?)(,\\s*(\\d+(-\\d+)?))*$`.
@@ -412,7 +412,7 @@ class AzureDocumentIntelligenceOCRConfig(BaseOCRConfig):
             raise ValueError("Document URL is required")
 
         # Build Azure DI request
-        data: Final[dict[str, Any]] = {}
+        data: Final[dict[str, str]] = {}
 
         # Check if it's a data URI (base64)
         if document_url.startswith("data:"):
diff --git a/litellm/llms/base_llm/guardrail_translation/base_translation.py b/litellm/llms/base_llm/guardrail_translation/base_translation.py
index 89ad67f0485..ace5af8124f 100644
--- a/litellm/llms/base_llm/guardrail_translation/base_translation.py
+++ b/litellm/llms/base_llm/guardrail_translation/base_translation.py
@@ -81,7 +81,7 @@ class BaseTranslation(ABC):
 
     @staticmethod
     def transform_user_api_key_dict_to_metadata(
-        user_api_key_dict: Any | None,
+        user_api_key_dict: Optional["UserAPIKeyAuth"],
     ) -> dict[str, object]:
         """
         Transform user_api_key_dict to a metadata dict with prefixed keys.
diff --git a/litellm/llms/bedrock/batches/transformation.py b/litellm/llms/bedrock/batches/transformation.py
index ae0f8c5935b..973388ca5bd 100644
--- a/litellm/llms/bedrock/batches/transformation.py
+++ b/litellm/llms/bedrock/batches/transformation.py
@@ -2,7 +2,7 @@ import os
 import re
 import time
 from collections.abc import Mapping
-from typing import TYPE_CHECKING, Any, Final, Literal, cast
+from typing import TYPE_CHECKING, Final, Literal, cast
 
 from httpx import Headers, Response
 from pydantic import TypeAdapter, ValidationError
@@ -170,7 +170,7 @@ class BedrockBatchesConfig(BaseAWSLLM, BaseBatchesConfig):
         create_batch_data: CreateBatchRequest,
         optional_params: dict,
         litellm_params: dict,
-    ) -> dict[str, Any]:
+    ) -> dict[str, object]:
         """
         Transform the batch creation request to Bedrock format.
 
@@ -354,7 +354,7 @@ class BedrockBatchesConfig(BaseAWSLLM, BaseBatchesConfig):
         )
 
     @staticmethod
-    def _get_openai_compatible_batch_metadata(metadata: Any) -> dict[str, str]:
+    def _get_openai_compatible_batch_metadata(metadata: object) -> dict[str, str]:
         """
         OpenAI Batch metadata only accepts string values.
         """
@@ -379,7 +379,7 @@ class BedrockBatchesConfig(BaseAWSLLM, BaseBatchesConfig):
         batch_id: str,
         optional_params: dict,
         litellm_params: dict,
-    ) -> dict[str, Any]:
+    ) -> dict[str, object]:
         """
         Transform batch retrieval request for Bedrock.
 
@@ -523,7 +523,7 @@ class BedrockBatchesConfig(BaseAWSLLM, BaseBatchesConfig):
             )
 
         # Enrich metadata with useful Bedrock fields
-        enriched_metadata_raw: Final[dict[str, Any]] = {
+        enriched_metadata_raw: Final[dict[str, object]] = {
             "jobName": response_data.get("jobName"),
             "clientRequestToken": response_data.get("clientRequestToken"),
             "modelId": response_data.get("modelId"),
diff --git a/litellm/llms/bedrock/chat/invoke_transformations/amazon_twelvelabs_pegasus_transformation.py b/litellm/llms/bedrock/chat/invoke_transformations/amazon_twelvelabs_pegasus_transformation.py
index d12c8aee48c..39cded4ed64 100644
--- a/litellm/llms/bedrock/chat/invoke_transformations/amazon_twelvelabs_pegasus_transformation.py
+++ b/litellm/llms/bedrock/chat/invoke_transformations/amazon_twelvelabs_pegasus_transformation.py
@@ -110,7 +110,7 @@ class AmazonTwelveLabsPegasusConfig(AmazonInvokeConfig, BaseConfig):
         headers: dict,
     ) -> dict:
         input_prompt: Final = self._convert_messages_to_prompt(messages=messages)
-        request_data: Final[dict[str, Any]] = {"inputPrompt": input_prompt}
+        request_data: Final[dict[str, object]] = {"inputPrompt": input_prompt}
 
         media_source: Final = self._build_media_source(optional_params)
         if media_source is not None:
diff --git a/litellm/llms/bytez/chat/transformation.py b/litellm/llms/bytez/chat/transformation.py
index d9a0c98b6db..7977db0f056 100644
--- a/litellm/llms/bytez/chat/transformation.py
+++ b/litellm/llms/bytez/chat/transformation.py
@@ -335,10 +335,10 @@ class BytezChatConfig(BaseConfig):
 
 
 class BytezCustomStreamWrapper(CustomStreamWrapper):
-    def chunk_creator(self, chunk: Any):
+    def chunk_creator(self, chunk: object):
         try:
             model_response: Final = self.model_response_creator()
-            response_obj: dict[str, Any] = {}
+            response_obj: dict[str, object] = {}
 
             response_obj = {
                 "text": chunk,
@@ -346,7 +346,7 @@ class BytezCustomStreamWrapper(CustomStreamWrapper):
                 "finish_reason": "",
             }
 
-            completion_obj: Final[dict[str, Any]] = {"content": chunk}
+            completion_obj: Final[dict[str, object]] = {"content": chunk}
 
             return self.return_processed_chunk_logic(
                 completion_obj=completion_obj,
diff --git a/litellm/llms/custom_httpx/aiohttp_handler.py b/litellm/llms/custom_httpx/aiohttp_handler.py
index 7035ce58ae1..0809ef5274f 100644
--- a/litellm/llms/custom_httpx/aiohttp_handler.py
+++ b/litellm/llms/custom_httpx/aiohttp_handler.py
@@ -1,5 +1,5 @@
 import ssl
-from collections.abc import Callable
+from collections.abc import AsyncIterable, Callable, Iterable
 from typing import TYPE_CHECKING, Any, Final, cast
 
 import aiohttp
@@ -212,7 +212,7 @@ class BaseLLMAIOHTTPHandler:
         litellm_params: dict,
         stream: bool = False,
         files: dict | None = None,
-        content: Any = None,
+        content: str | bytes | Iterable[bytes] | AsyncIterable[bytes] | None = None,
         params: dict | None = None,
     ) -> httpx.Response:
         max_retry_on_unprocessable_entity_error: Final = provider_config.max_retry_on_unprocessable_entity_error
diff --git a/litellm/llms/deprecated_providers/aleph_alpha.py b/litellm/llms/deprecated_providers/aleph_alpha.py
index 4a29549b6aa..2ad9ce4edc8 100644
--- a/litellm/llms/deprecated_providers/aleph_alpha.py
+++ b/litellm/llms/deprecated_providers/aleph_alpha.py
@@ -146,7 +146,7 @@ class AlephAlphaConfig:
                 setattr(self.__class__, key, value)
 
     @classmethod
-    def get_config(cls):
+    def get_config(cls) -> dict[str, object]:
         return {
             k: v
             for k, v in cls.__dict__.items()
diff --git a/litellm/llms/lemonade/chat/transformation.py b/litellm/llms/lemonade/chat/transformation.py
index 553478aec16..c01ad2a0edd 100644
--- a/litellm/llms/lemonade/chat/transformation.py
+++ b/litellm/llms/lemonade/chat/transformation.py
@@ -170,7 +170,7 @@ class LemonadeChatConfig(OpenAILikeChatConfig):
         model: str,
         api_base: str | None = None,
         api_key: str | None = None,
-    ) -> Any:
+    ) -> dict[str, object]:
         if model.startswith("lemonade/"):
             model = model.split("/", 1)[1]
 
diff --git a/litellm/llms/openai/image_edit/transformation.py b/litellm/llms/openai/image_edit/transformation.py
index f55084adbde..d54522597a0 100644
--- a/litellm/llms/openai/image_edit/transformation.py
+++ b/litellm/llms/openai/image_edit/transformation.py
@@ -66,7 +66,7 @@ class OpenAIImageEditConfig(BaseImageEditConfig):
     def _add_image_to_files(
         self,
         files_list: list[tuple[str, Any]],
-        image: Any,
+        image: object,
         field_name: str,
     ) -> None:
         """Add an image to the files list with appropriate content type"""
diff --git a/litellm/llms/runwayml/text_to_speech/transformation.py b/litellm/llms/runwayml/text_to_speech/transformation.py
index 19e6d8ff494..6769accc1d6 100644
--- a/litellm/llms/runwayml/text_to_speech/transformation.py
+++ b/litellm/llms/runwayml/text_to_speech/transformation.py
@@ -78,7 +78,7 @@ class RunwayMLTextToSpeechConfig(BaseTextToSpeechConfig):
         aspeech: bool,
         api_base: str | None,
         api_key: str | None,
-        **kwargs: Any,
+        **kwargs: object,
     ) -> Union[
         "HttpxBinaryResponseContent",
         Coroutine[object, object, "HttpxBinaryResponseContent"],
diff --git a/litellm/llms/vertex_ai/files/transformation.py b/litellm/llms/vertex_ai/files/transformation.py
index 85ec2911464..80d32289c94 100644
--- a/litellm/llms/vertex_ai/files/transformation.py
+++ b/litellm/llms/vertex_ai/files/transformation.py
@@ -651,7 +651,7 @@ def _openai_batch_jsonl_entry_to_vertex_embeddings_rows(
 
 def _openai_batch_jsonl_entry_to_vertex_rows(
     openai_entry: dict[str, Any],
-    map_openai_to_vertex_params: Callable[[dict[str, Any]], dict[str, Any]],
+    map_openai_to_vertex_params: Callable[[dict[str, Any]], dict[str, object]],
 ) -> tuple[Mapping[str, object], ...]:
     """
     Transforms a single OpenAI JSONL batch entry into the Vertex rows it maps to.
@@ -774,7 +774,7 @@ class _OpenAIToVertexBatchUploadStream(BaseFileUploadStream):
     def __init__(
         self,
         openai_file_content: FileTypes,
-        map_openai_to_vertex_params: Callable[[dict[str, Any]], dict[str, Any]],
+        map_openai_to_vertex_params: Callable[[dict[str, Any]], dict[str, object]],
     ) -> None:
         self._openai_file_content = openai_file_content
         self._map_openai_to_vertex_params = map_openai_to_vertex_params
@@ -948,7 +948,7 @@ class VertexAIFilesConfig(VertexBase, BaseFilesConfig):
     def _map_openai_to_vertex_params(
         self,
         openai_request_body: dict[str, Any],
-    ) -> dict[str, Any]:
+    ) -> dict[str, object]:
         """
         wrapper to call VertexGeminiConfig.map_openai_params
         """
diff --git a/litellm/llms/vertex_ai/gemini_embeddings/batch_embed_content_handler.py b/litellm/llms/vertex_ai/gemini_embeddings/batch_embed_content_handler.py
index f81d4ca777e..c6ac87d646b 100644
--- a/litellm/llms/vertex_ai/gemini_embeddings/batch_embed_content_handler.py
+++ b/litellm/llms/vertex_ai/gemini_embeddings/batch_embed_content_handler.py
@@ -3,7 +3,7 @@ Google AI Studio /batchEmbedContents Embeddings Endpoint
 """
 
 import json
-from typing import TYPE_CHECKING, Any, Final, Literal
+from typing import TYPE_CHECKING, Final, Literal
 
 import httpx
 
@@ -210,7 +210,7 @@ class GoogleBatchEmbeddings(VertexLLM):
             )
 
         ### TRANSFORMATION (sync path) ###
-        request_data: Any
+        request_data: VertexAIBatchEmbeddingsRequestBody | dict[str, object]
         if use_embed_content:
             resolved_files = {}
             if api_key:
diff --git a/litellm/llms/vertex_ai/vertex_ai_non_gemini.py b/litellm/llms/vertex_ai/vertex_ai_non_gemini.py
index 1c582c7c376..a7a1ea8d88d 100644
--- a/litellm/llms/vertex_ai/vertex_ai_non_gemini.py
+++ b/litellm/llms/vertex_ai/vertex_ai_non_gemini.py
@@ -64,7 +64,7 @@ def _get_client_from_cache(client_cache_key: str):
     return litellm.in_memory_llm_clients_cache.get_cache(client_cache_key)
 
 
-def _set_client_in_cache(client_cache_key: str, vertex_llm_model: Any):
+def _set_client_in_cache(client_cache_key: str, vertex_llm_model: object):
     litellm.in_memory_llm_clients_cache.set_cache(
         key=client_cache_key,
         value=vertex_llm_model,
diff --git a/litellm/llms/watsonx/audio_transcription/transformation.py b/litellm/llms/watsonx/audio_transcription/transformation.py
index 7d1aba63428..2169b9bf49a 100644
--- a/litellm/llms/watsonx/audio_transcription/transformation.py
+++ b/litellm/llms/watsonx/audio_transcription/transformation.py
@@ -4,7 +4,7 @@ Translates from OpenAI's `/v1/audio/transcriptions` to IBM WatsonX's `/ml/v1/aud
 WatsonX follows the OpenAI spec for audio transcription.
 """
 
-from typing import Any, Final
+from typing import Final
 
 from httpx import Response
 
@@ -124,7 +124,7 @@ class IBMWatsonXAudioTranscriptionConfig(IBMWatsonXMixin, OpenAIWhisperAudioTran
         }
 
         # Convert TypedDict to regular dict for AudioTranscriptionRequestData
-        form_data_dict: Final[dict[str, Any]] = dict(form_data)
+        form_data_dict: Final[dict[str, object]] = dict(form_data)
 
         return AudioTranscriptionRequestData(data=form_data_dict, files=files)
 
diff --git a/litellm/proxy/a2a/agent_card.py b/litellm/proxy/a2a/agent_card.py
index 5bec5158bcc..3718359fbb6 100644
--- a/litellm/proxy/a2a/agent_card.py
+++ b/litellm/proxy/a2a/agent_card.py
@@ -10,7 +10,7 @@ and uses LiteLLM auth.
 import re
 from collections.abc import Mapping
 from copy import deepcopy
-from typing import Any, Final, Literal
+from typing import Final, Literal
 
 SupportedA2AVersion = Literal["0.3", "1.0"]
 
@@ -44,7 +44,7 @@ def normalize_protocol_version(version: object) -> SupportedA2AVersion | None:
     return next((supported for supported in SUPPORTED_A2A_PROTOCOL_VERSIONS if supported == major_minor), None)
 
 
-def resolve_served_protocol_version(card: Mapping[str, Any] | None) -> str:
+def resolve_served_protocol_version(card: Mapping[str, object] | None) -> str:
     """Return the validated protocol version an agent card pins, else the default."""
     normalized: Final = normalize_protocol_version(card.get("protocolVersion") if card else None)
     return normalized if normalized is not None else LITELLM_A2A_PROTOCOL_VERSION
@@ -53,7 +53,7 @@ def resolve_served_protocol_version(card: Mapping[str, Any] | None) -> str:
 # Security scheme exposed by the LiteLLM-fronted agent card. Always replaces
 # whatever upstream advertised — the client must authenticate to the proxy,
 # not the upstream agent.
-LITELLM_SECURITY_SCHEMES: Final[dict[str, dict[str, Any]]] = {
+LITELLM_SECURITY_SCHEMES: Final[dict[str, dict[str, str]]] = {
     "LiteLLMKey": {
         "type": "http",
         "scheme": "bearer",
@@ -112,7 +112,7 @@ _ALLOWED_TOP_LEVEL_KEYS: Final = {
     "url",
 }
 
-_DEFAULT_SKILLS: Final[list[dict[str, Any]]] = [
+_DEFAULT_SKILLS: Final[list[dict[str, str | list[str]]]] = [
     {
         "id": "chat",
         "name": "Chat",
@@ -129,7 +129,7 @@ _DEFAULT_MODES: Final[list[str]] = ["text"]
 _DEFAULT_AGENT_VERSION: Final = "1.0.0"
 
 
-def _filter_capabilities(upstream_capabilities: Any) -> dict[str, Any]:
+def _filter_capabilities(upstream_capabilities: object) -> dict[str, object]:
     """Return a capabilities dict containing only allowlisted, truthy keys."""
     if not isinstance(upstream_capabilities, dict):
         return {}
@@ -143,13 +143,13 @@ def _default_litellm_provider(proxy_base_url: str) -> dict[str, str]:
 
 
 def merge_agent_card(
-    upstream_card: Mapping[str, Any] | None,
+    upstream_card: Mapping[str, object] | None,
     *,
     proxy_url: str,
     proxy_base_url: str,
     name: str | None = None,
     description: str | None = None,
-) -> dict[str, Any]:
+) -> dict[str, object]:
     """
     Build the LiteLLM-fronted agent card.
 
@@ -169,7 +169,7 @@ def merge_agent_card(
         A dict suitable for serving as the proxy's agent card. Only keys in
         the v1.0 AgentCard schema (plus ``supportedInterfaces``) are emitted.
     """
-    base: Final[dict[str, Any]] = deepcopy(dict(upstream_card)) if upstream_card else {}
+    base: Final[dict[str, object]] = deepcopy(dict(upstream_card)) if upstream_card else {}
 
     # Keep the upstream ``url`` on the stored card: the runtime A2A
     # invocation path reads it from ``agent_card_params`` to know where to
diff --git a/litellm/proxy/agent_endpoints/a2a_endpoints.py b/litellm/proxy/agent_endpoints/a2a_endpoints.py
index 834c16ba6dc..faf3e98a3a7 100644
--- a/litellm/proxy/agent_endpoints/a2a_endpoints.py
+++ b/litellm/proxy/agent_endpoints/a2a_endpoints.py
@@ -880,7 +880,7 @@ async def invoke_agent_a2a(
                     logging_obj._enqueue_deferred_logging = None
                     _enqueue_fn()
 
-            response_dict: Final[dict[str, Any]] = (
+            response_dict: Final[dict[str, object]] = (
                 response.model_dump(mode="json", exclude_none=True)
                 if hasattr(response, "model_dump")
                 else response
diff --git a/litellm/proxy/client/credentials.py b/litellm/proxy/client/credentials.py
index a9bff67b1c5..d9edecd2eb7 100644
--- a/litellm/proxy/client/credentials.py
+++ b/litellm/proxy/client/credentials.py
@@ -1,3 +1,4 @@
+from collections.abc import Mapping
 from typing import Any, Final
 
 import requests
@@ -69,8 +70,8 @@ class CredentialsManagementClient:
     def create(
         self,
         credential_name: str,
-        credential_info: dict[str, Any],
-        credential_values: dict[str, Any],
+        credential_info: Mapping[str, object],
+        credential_values: Mapping[str, object],
         return_request: bool = False,
     ) -> dict[str, Any] | requests.Request:
         """
diff --git a/litellm/proxy/guardrails/guardrail_hooks/guardrails_ai/guardrails_ai.py b/litellm/proxy/guardrails/guardrail_hooks/guardrails_ai/guardrails_ai.py
index 47324471650..18451df574f 100644
--- a/litellm/proxy/guardrails/guardrail_hooks/guardrails_ai/guardrails_ai.py
+++ b/litellm/proxy/guardrails/guardrail_hooks/guardrails_ai/guardrails_ai.py
@@ -7,7 +7,7 @@
 
 import json
 import os
-from typing import TYPE_CHECKING, Any, Final, Literal, TypedDict
+from typing import TYPE_CHECKING, Final, Literal, TypedDict
 
 from fastapi import HTTPException
 
@@ -181,7 +181,7 @@ class GuardrailsAI(CustomGuardrail):
     ):  # raise exception if invalid, return a str for the user to receive - if rejected, or return a modified dictionary for passing into litellm
         return await self.process_input(data=data, call_type=call_type)
 
-    async def async_logging_hook(self, kwargs: dict, result: Any, call_type: str) -> tuple[dict, Any]:
+    async def async_logging_hook(self, kwargs: dict, result: object, call_type: str) -> tuple[dict, object]:
         if call_type == "acompletion" or call_type == "completion":
             kwargs = await self.process_input(data=kwargs, call_type=call_type)
 
diff --git a/litellm/proxy/guardrails/guardrail_hooks/singulr/singulr.py b/litellm/proxy/guardrails/guardrail_hooks/singulr/singulr.py
index 06d4b39f5f6..bd5b18e368d 100644
--- a/litellm/proxy/guardrails/guardrail_hooks/singulr/singulr.py
+++ b/litellm/proxy/guardrails/guardrail_hooks/singulr/singulr.py
@@ -24,7 +24,7 @@ from litellm.llms.custom_httpx.http_handler import (
 )
 from litellm.proxy._types import UserAPIKeyAuth
 from litellm.types.guardrails import GuardrailEventHooks
-from litellm.types.llms.openai import AllMessageValues
+from litellm.types.llms.openai import AllMessageValues, ChatCompletionToolCallChunk
 from litellm.types.proxy.guardrails.guardrail_hooks.base import (
     GuardrailConfigModel,
 )
@@ -36,7 +36,7 @@ from litellm.types.proxy.guardrails.guardrail_hooks.singulr import (
     ToolCall,
     ToolCallFunction,
 )
-from litellm.types.utils import CallTypes, GenericGuardrailAPIInputs
+from litellm.types.utils import CallTypes, ChatCompletionMessageToolCall, GenericGuardrailAPIInputs
 
 _DEFAULT_API_BASE: Final = "http://localhost:8003"
 _GUARD_ENDPOINT: Final = "/api/v1/ai-gateway/litellm-v2"
@@ -339,7 +339,7 @@ class SingulrGuardrail(CustomGuardrail):
         return inputs
 
     @staticmethod
-    def _build_tool_call(tool_call: Mapping[str, Any]) -> "ToolCall | None":
+    def _build_tool_call(tool_call: ChatCompletionToolCallChunk | ChatCompletionMessageToolCall) -> "ToolCall | None":
         tool_call_id: Final = tool_call.get("id")
         fun: Final = tool_call.get("function")
         if not tool_call_id or not fun:
diff --git a/litellm/proxy/management_helpers/object_permission_utils.py b/litellm/proxy/management_helpers/object_permission_utils.py
index daab38d3662..61e432daa16 100644
--- a/litellm/proxy/management_helpers/object_permission_utils.py
+++ b/litellm/proxy/management_helpers/object_permission_utils.py
@@ -8,7 +8,7 @@ from collections.abc import Mapping, Sequence
 from collections.abc import Set as AbstractSet
 from dataclasses import dataclass
 from types import MappingProxyType
-from typing import TYPE_CHECKING, Any, Final, Optional
+from typing import TYPE_CHECKING, Final, Optional
 
 from fastapi import HTTPException, status
 from pydantic import TypeAdapter
@@ -156,12 +156,13 @@ async def handle_update_object_permission_common(
     if prisma_client is None:
         raise ValueError("Prisma client not found")
 
-    new_object_permission: dict | str | None = data_json.pop("object_permission", None)
-    if new_object_permission is None:
+    raw_object_permission: Final[dict | str | None] = data_json.pop("object_permission", None)
+    if raw_object_permission is None:
         return None
 
-    if isinstance(new_object_permission, str):
-        new_object_permission = json.loads(new_object_permission)
+    new_object_permission: Final[object] = (
+        json.loads(raw_object_permission) if isinstance(raw_object_permission, str) else raw_object_permission
+    )
 
     upsert: Final = await prepare_object_permission_upsert(
         new_object_permission=new_object_permission if isinstance(new_object_permission, dict) else {},
@@ -230,7 +231,7 @@ def _dedupe_preserving_order(values: list[str]) -> list[str]:
     return result
 
 
-def _mcp_server_identifier_matches(server: Any, identifier: str) -> bool:
+def _mcp_server_identifier_matches(server: object, identifier: str) -> bool:
     return identifier in {
         getattr(server, "server_id", None),
         getattr(server, "alias", None),
diff --git a/litellm/proxy/policy_engine/pipeline_executor.py b/litellm/proxy/policy_engine/pipeline_executor.py
index 0b81e7af84d..e9d23436b59 100644
--- a/litellm/proxy/policy_engine/pipeline_executor.py
+++ b/litellm/proxy/policy_engine/pipeline_executor.py
@@ -8,7 +8,7 @@ pass/fail actions (allow, block, next, modify_response) and data forwarding.
 import copy
 import time
 from collections.abc import Callable, Mapping, Sequence
-from typing import TYPE_CHECKING, Any, Final, Literal, TypeVar
+from typing import TYPE_CHECKING, Final, Literal, TypeVar
 
 from pydantic import BaseModel
 
@@ -314,11 +314,11 @@ class PipelineExecutor:
         steps: list[PipelineStep],
         mode: str,
         data: dict,
-        user_api_key_dict: Any,
+        user_api_key_dict: "UserAPIKeyAuth",
         call_type: str,
         policy_name: str,
         raw_request_snapshot: dict | None = None,  # mutable-ok: same request-payload shape as data
-        streaming_chunks: list[Any] | None = None,  # mutable-ok: shared buffered-stream chunks, read per step
+        streaming_chunks: list[object] | None = None,  # mutable-ok: shared buffered-stream chunks, read per step
         endpoint_translation: "BaseTranslation | None" = None,
     ) -> PipelineExecutionResult:
         """
@@ -490,10 +490,10 @@ class PipelineExecutor:
         step: PipelineStep,
         mode: str,
         data: dict,
-        user_api_key_dict: Any,
+        user_api_key_dict: "UserAPIKeyAuth",
         call_type: str,
         raw_request_snapshot: dict | None = None,  # mutable-ok: same request-payload shape as data
-        streaming_chunks: list[Any] | None = None,  # mutable-ok: shared buffered-stream chunks, read per step
+        streaming_chunks: list[object] | None = None,  # mutable-ok: shared buffered-stream chunks, read per step
         endpoint_translation: "BaseTranslation | None" = None,
     ) -> tuple[
         Literal["pass", "fail", "error"],
@@ -722,7 +722,7 @@ def _extract_error_message(e: Exception) -> str:
     if isinstance(e, ModifyResponseException):
         return str(e)
     if HTTPException is not None and isinstance(e, HTTPException):
-        detail: Final = getattr(e, "detail", None)
+        detail: Final[object] = getattr(e, "detail", None)
         if detail:
             return str(detail)
     return str(e)
diff --git a/litellm/router_strategy/adaptive_router/hooks.py b/litellm/router_strategy/adaptive_router/hooks.py
index 709910753f2..c4a2eae1ef9 100644
--- a/litellm/router_strategy/adaptive_router/hooks.py
+++ b/litellm/router_strategy/adaptive_router/hooks.py
@@ -86,7 +86,7 @@ def _resolve_session_key(kwargs: dict[str, Any]) -> str | None:
     return hashlib.sha256(payload.encode("utf-8")).hexdigest()
 
 
-def _last_user_content(messages: list[dict[str, Any]] | None) -> str | None:
+def _last_user_content(messages: Sequence[Mapping[str, object]] | None) -> str | None:
     if not messages:
         return None
     for msg in reversed(messages):
diff --git a/litellm/router_strategy/auto_router/auto_router.py b/litellm/router_strategy/auto_router/auto_router.py
index d08afa8c1f6..250201a46d1 100644
--- a/litellm/router_strategy/auto_router/auto_router.py
+++ b/litellm/router_strategy/auto_router/auto_router.py
@@ -3,7 +3,7 @@ Auto-Routing Strategy that works with a Semantic Router Config
 """
 
 import asyncio
-from collections.abc import Mapping
+from collections.abc import Mapping, Sequence
 from typing import TYPE_CHECKING, Any, Final, Optional
 
 from pydantic import BaseModel, ConfigDict
@@ -158,7 +158,7 @@ class AutoRouter(CustomLogger):
         return await asyncio.shield(build_task)
 
     @staticmethod
-    def _extract_text_from_messages(messages: list[dict[str, Any]]) -> str:
+    def _extract_text_from_messages(messages: Sequence[Mapping[str, object]]) -> str:
         """
         Extract text content from the last user message for routing.
 

From dbf566873ea32a82a9915876615dad67febd2e57 Mon Sep 17 00:00:00 2001
From: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Date: Sun, 20 Sep 2026 10:28:06 +0000
Subject: [PATCH 09/13] refactor(types): keep agentops preset imports optional

Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
---
 litellm/integrations/otel/presets/agentops.py | 7 ++-----
 1 file changed, 2 insertions(+), 5 deletions(-)

diff --git a/litellm/integrations/otel/presets/agentops.py b/litellm/integrations/otel/presets/agentops.py
index 58123656caa..965213f2ee4 100644
--- a/litellm/integrations/otel/presets/agentops.py
+++ b/litellm/integrations/otel/presets/agentops.py
@@ -9,12 +9,9 @@ this preset registers a custom exporter (``kind="agentops"``) that mints the JWT
 worker thread, off any event loop — and caches it for the process lifetime.
 """
 
-from collections.abc import Sequence
 from typing import Any, Final
 
 import httpx
-from opentelemetry.sdk.trace import ReadableSpan
-from opentelemetry.sdk.trace.export import SpanExporter, SpanExportResult
 from pydantic import Field
 from pydantic_settings import BaseSettings, SettingsConfigDict
 
@@ -74,7 +71,7 @@ def agentops_preset(
     )
 
 
-def _build_agentops_exporter(spec: ExporterSpec) -> SpanExporter:
+def _build_agentops_exporter(spec: ExporterSpec) -> Any:
     """Factory for the ``agentops`` exporter kind: a lazy-auth OTLP/HTTP exporter."""
     from opentelemetry.exporter.otlp.proto.http.trace_exporter import (
         OTLPSpanExporter,
@@ -109,7 +106,7 @@ def _build_agentops_exporter(spec: ExporterSpec) -> SpanExporter:
             except Exception as e:
                 verbose_logger.debug("AgentOps JWT fetch failed: %s", e)
 
-        def export(self, spans: Sequence[ReadableSpan]) -> SpanExportResult:
+        def export(self, spans: Any) -> Any:
             self._ensure_authenticated()
             return super().export(spans)
 

From e0b92b2b257bd26650a5887343cfc0de3127129d Mon Sep 17 00:00:00 2001
From: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Date: Sun, 20 Sep 2026 11:07:38 +0000
Subject: [PATCH 10/13] refactor(types): keep a2a response_dict annotation as
 it was

Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
---
 litellm/proxy/agent_endpoints/a2a_endpoints.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/litellm/proxy/agent_endpoints/a2a_endpoints.py b/litellm/proxy/agent_endpoints/a2a_endpoints.py
index faf3e98a3a7..834c16ba6dc 100644
--- a/litellm/proxy/agent_endpoints/a2a_endpoints.py
+++ b/litellm/proxy/agent_endpoints/a2a_endpoints.py
@@ -880,7 +880,7 @@ async def invoke_agent_a2a(
                     logging_obj._enqueue_deferred_logging = None
                     _enqueue_fn()
 
-            response_dict: Final[dict[str, object]] = (
+            response_dict: Final[dict[str, Any]] = (
                 response.model_dump(mode="json", exclude_none=True)
                 if hasattr(response, "model_dump")
                 else response

From 6ba4c2e3399bc16a9996c379c020d1b726f8247a Mon Sep 17 00:00:00 2001
From: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Date: Sun, 20 Sep 2026 11:40:11 +0000
Subject: [PATCH 11/13] refactor(types): keep guardrail metadata helper
 accepting dicts

Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
---
 litellm/llms/base_llm/guardrail_translation/base_translation.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/litellm/llms/base_llm/guardrail_translation/base_translation.py b/litellm/llms/base_llm/guardrail_translation/base_translation.py
index ace5af8124f..89ad67f0485 100644
--- a/litellm/llms/base_llm/guardrail_translation/base_translation.py
+++ b/litellm/llms/base_llm/guardrail_translation/base_translation.py
@@ -81,7 +81,7 @@ class BaseTranslation(ABC):
 
     @staticmethod
     def transform_user_api_key_dict_to_metadata(
-        user_api_key_dict: Optional["UserAPIKeyAuth"],
+        user_api_key_dict: Any | None,
     ) -> dict[str, object]:
         """
         Transform user_api_key_dict to a metadata dict with prefixed keys.

From e4a58ef91acdebe2a25c66f7dbdf7d4bdc338fd7 Mon Sep 17 00:00:00 2001
From: yuneng <yuneng@berri.ai>
Date: Sun, 20 Sep 2026 11:50:59 +0000
Subject: [PATCH 12/13] test(unit): make every tests/unit directory a package
 so pytest collection is unique

Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
---
 tests/unit/__init__.py                               |  0
 tests/unit/integrations/__init__.py                  |  0
 tests/unit/integrations/levo/__init__.py             |  0
 tests/unit/integrations/litellm_agent/__init__.py    |  0
 tests/unit/integrations/mavvrik_focus/__init__.py    |  0
 tests/unit/integrations/opik/__init__.py             |  0
 tests/unit/integrations/pointfive/__init__.py        |  0
 .../vector_store_integrations/__init__.py            |  0
 tests/unit/litellm_core_utils/__init__.py            |  0
 .../unit/litellm_core_utils/audio_utils/__init__.py  |  0
 .../llm_response_utils/__init__.py                   |  0
 tests/unit/llms/__init__.py                          |  0
 tests/unit/llms/a2a/__init__.py                      |  0
 tests/unit/llms/a2a/chat/__init__.py                 |  0
 .../llms/a2a/chat/guardrail_translation/__init__.py  |  0
 tests/unit/llms/anthropic/__init__.py                |  0
 tests/unit/llms/anthropic/batches/__init__.py        |  0
 tests/unit/llms/base_llm/__init__.py                 |  0
 tests/unit/llms/base_llm/batches/__init__.py         |  0
 tests/unit/llms/base_llm/realtime/__init__.py        |  0
 tests/unit/llms/baseten/__init__.py                  |  0
 tests/unit/llms/baseten/chat/__init__.py             |  0
 tests/unit/llms/bedrock/__init__.py                  |  0
 tests/unit/llms/bedrock/chat/__init__.py             |  0
 tests/unit/llms/bedrock/chat/agentcore/__init__.py   |  0
 .../bedrock/chat/invoke_transformations/__init__.py  |  0
 tests/unit/llms/bedrock/chat/mantle/__init__.py      |  0
 tests/unit/llms/bedrock/count_tokens/__init__.py     |  0
 tests/unit/llms/bedrock/files/__init__.py            |  0
 tests/unit/llms/bedrock/image/__init__.py            |  0
 tests/unit/llms/bedrock/image_edit/__init__.py       |  0
 tests/unit/llms/bedrock/invoke_agent/__init__.py     |  0
 tests/unit/llms/bedrock/passthrough/__init__.py      |  0
 .../passthrough/guardrail_translation/__init__.py    |  0
 tests/unit/llms/bedrock/realtime/__init__.py         |  0
 tests/unit/llms/bedrock/rerank/__init__.py           |  0
 tests/unit/llms/bedrock/vector_stores/__init__.py    |  0
 tests/unit/llms/bedrock_mantle/__init__.py           |  0
 .../unit/llms/bedrock_mantle/passthrough/__init__.py |  0
 tests/unit/llms/black_forest_labs/__init__.py        |  0
 .../llms/black_forest_labs/image_edit/__init__.py    |  0
 .../black_forest_labs/image_generation/__init__.py   |  0
 tests/unit/llms/bytez/__init__.py                    |  0
 tests/unit/llms/bytez/chat/__init__.py               |  0
 tests/unit/llms/cerebras/__init__.py                 |  0
 tests/unit/llms/chat/__init__.py                     |  0
 tests/unit/llms/chatgpt/__init__.py                  |  0
 tests/unit/llms/chatgpt/chat/__init__.py             |  0
 tests/unit/llms/chatgpt/responses/__init__.py        |  0
 tests/unit/llms/cloudflare/__init__.py               |  0
 tests/unit/llms/cohere/__init__.py                   |  0
 tests/unit/llms/cohere/chat/__init__.py              |  0
 tests/unit/llms/cohere/embed/__init__.py             |  0
 tests/unit/llms/cohere/ocr/__init__.py               |  0
 tests/unit/llms/cohere/rerank/__init__.py            |  0
 tests/unit/llms/crusoe/__init__.py                   |  0
 tests/unit/llms/databricks/__init__.py               |  0
 tests/unit/llms/databricks/chat/__init__.py          |  0
 tests/unit/llms/databricks/responses/__init__.py     |  0
 tests/unit/llms/datarobot/__init__.py                |  0
 tests/unit/llms/datarobot/chat/__init__.py           |  0
 tests/unit/llms/deepseek/__init__.py                 |  0
 tests/unit/llms/deepseek/chat/__init__.py            |  0
 tests/unit/llms/deepseek/messages/__init__.py        |  0
 tests/unit/llms/docker_model_runner/__init__.py      |  0
 tests/unit/llms/elevenlabs/__init__.py               |  0
 tests/unit/llms/fastcrw/__init__.py                  |  0
 tests/unit/llms/fastcrw/search/__init__.py           |  0
 tests/unit/llms/fireworks_ai/__init__.py             |  0
 tests/unit/llms/fireworks_ai/chat/__init__.py        |  0
 tests/unit/llms/fireworks_ai/rerank/__init__.py      |  0
 tests/unit/llms/fireworks_ai/responses/__init__.py   |  0
 tests/unit/llms/gemini/__init__.py                   |  0
 .../unit/llms/gemini/audio_transcription/__init__.py |  0
 tests/unit/llms/gemini/files/__init__.py             |  0
 tests/unit/llms/gemini/google_genai/__init__.py      |  0
 .../google_genai/guardrail_translation/__init__.py   |  0
 tests/unit/llms/gemini/image_edit/__init__.py        |  0
 tests/unit/llms/gemini/realtime/__init__.py          |  0
 tests/unit/llms/gemini/videos/__init__.py            |  0
 tests/unit/llms/gigachat/__init__.py                 |  0
 tests/unit/llms/gigachat/chat/__init__.py            |  0
 tests/unit/llms/gigachat/embedding/__init__.py       |  0
 tests/unit/llms/gigachat/passthrough/__init__.py     |  0
 tests/unit/llms/github_copilot/__init__.py           |  0
 tests/unit/llms/github_copilot/embedding/__init__.py |  0
 tests/unit/llms/github_copilot/messages/__init__.py  |  0
 tests/unit/llms/github_copilot/responses/__init__.py |  0
 tests/unit/llms/gradient_ai/__init__.py              |  0
 tests/unit/llms/gradient_ai/chat/__init__.py         |  0
 tests/unit/llms/groq/__init__.py                     |  0
 tests/unit/llms/groq/chat/__init__.py                |  0
 tests/unit/llms/hosted_vllm/__init__.py              |  0
 tests/unit/llms/hosted_vllm/chat/__init__.py         |  0
 tests/unit/llms/hosted_vllm/embedding/__init__.py    |  0
 tests/unit/llms/hosted_vllm/image_edit/__init__.py   |  0
 tests/unit/llms/hosted_vllm/responses/__init__.py    |  0
 tests/unit/llms/hosted_vllm/videos/__init__.py       |  0
 tests/unit/llms/huggingface/__init__.py              |  0
 tests/unit/llms/huggingface/rerank/__init__.py       |  0
 tests/unit/llms/inception/__init__.py                |  0
 tests/unit/llms/jina_ai/__init__.py                  |  0
 tests/unit/llms/jina_ai/embedding/__init__.py        |  0
 tests/unit/llms/langflow/__init__.py                 |  0
 tests/unit/llms/langflow/chat/__init__.py            |  0
 tests/unit/llms/litellm_proxy/__init__.py            |  0
 tests/unit/llms/litellm_proxy/chat/__init__.py       |  0
 tests/unit/llms/litellm_proxy/skills/__init__.py     |  0
 tests/unit/llms/llamafile/__init__.py                |  0
 tests/unit/llms/llamafile/chat/__init__.py           |  0
 tests/unit/llms/meta/__init__.py                     |  0
 tests/unit/llms/meta/realtime/__init__.py            |  0
 tests/unit/llms/meta_llama/__init__.py               |  0
 tests/unit/llms/mistral/audio_speech/__init__.py     |  0
 tests/unit/llms/modelscope/__init__.py               |  0
 .../llms/modelscope/image_generation/__init__.py     |  0
 tests/unit/llms/mongodb/__init__.py                  |  0
 tests/unit/llms/mongodb/vector_stores/__init__.py    |  0
 tests/unit/llms/moonshot/__init__.py                 |  0
 tests/unit/llms/neosantara/__init__.py               |  0
 tests/unit/llms/nimble/__init__.py                   |  0
 tests/unit/llms/nimble/search/__init__.py            |  0
 tests/unit/llms/novita/__init__.py                   |  0
 tests/unit/llms/novita/chat/__init__.py              |  0
 tests/unit/llms/nscale/__init__.py                   |  0
 tests/unit/llms/nscale/chat/__init__.py              |  0
 tests/unit/llms/nvidia_nim/__init__.py               |  0
 tests/unit/llms/nvidia_nim/passthrough/__init__.py   |  0
 tests/unit/llms/nvidia_nim/rerank/__init__.py        |  0
 tests/unit/llms/nvidia_riva/__init__.py              |  0
 .../llms/nvidia_riva/audio_transcription/__init__.py |  0
 tests/unit/llms/oci/__init__.py                      |  0
 tests/unit/llms/oci/chat/__init__.py                 |  0
 tests/unit/llms/oci/embed/__init__.py                |  0
 tests/unit/llms/ocr/__init__.py                      |  0
 .../unit/llms/ocr/guardrail_translation/__init__.py  |  0
 tests/unit/llms/oobabooga/__init__.py                |  0
 tests/unit/llms/oobabooga/chat/__init__.py           |  0
 tests/unit/llms/openai/__init__.py                   |  0
 tests/unit/llms/openai/chat/__init__.py              |  0
 .../openai/chat/guardrail_translation/__init__.py    |  0
 tests/unit/llms/openai/completion/__init__.py        |  0
 tests/unit/test_package_layout.py                    | 12 ++++++++++++
 143 files changed, 12 insertions(+)
 create mode 100644 tests/unit/__init__.py
 create mode 100644 tests/unit/integrations/__init__.py
 create mode 100644 tests/unit/integrations/levo/__init__.py
 create mode 100644 tests/unit/integrations/litellm_agent/__init__.py
 create mode 100644 tests/unit/integrations/mavvrik_focus/__init__.py
 create mode 100644 tests/unit/integrations/opik/__init__.py
 create mode 100644 tests/unit/integrations/pointfive/__init__.py
 create mode 100644 tests/unit/integrations/vector_store_integrations/__init__.py
 create mode 100644 tests/unit/litellm_core_utils/__init__.py
 create mode 100644 tests/unit/litellm_core_utils/audio_utils/__init__.py
 create mode 100644 tests/unit/litellm_core_utils/llm_response_utils/__init__.py
 create mode 100644 tests/unit/llms/__init__.py
 create mode 100644 tests/unit/llms/a2a/__init__.py
 create mode 100644 tests/unit/llms/a2a/chat/__init__.py
 create mode 100644 tests/unit/llms/a2a/chat/guardrail_translation/__init__.py
 create mode 100644 tests/unit/llms/anthropic/__init__.py
 create mode 100644 tests/unit/llms/anthropic/batches/__init__.py
 create mode 100644 tests/unit/llms/base_llm/__init__.py
 create mode 100644 tests/unit/llms/base_llm/batches/__init__.py
 create mode 100644 tests/unit/llms/base_llm/realtime/__init__.py
 create mode 100644 tests/unit/llms/baseten/__init__.py
 create mode 100644 tests/unit/llms/baseten/chat/__init__.py
 create mode 100644 tests/unit/llms/bedrock/__init__.py
 create mode 100644 tests/unit/llms/bedrock/chat/__init__.py
 create mode 100644 tests/unit/llms/bedrock/chat/agentcore/__init__.py
 create mode 100644 tests/unit/llms/bedrock/chat/invoke_transformations/__init__.py
 create mode 100644 tests/unit/llms/bedrock/chat/mantle/__init__.py
 create mode 100644 tests/unit/llms/bedrock/count_tokens/__init__.py
 create mode 100644 tests/unit/llms/bedrock/files/__init__.py
 create mode 100644 tests/unit/llms/bedrock/image/__init__.py
 create mode 100644 tests/unit/llms/bedrock/image_edit/__init__.py
 create mode 100644 tests/unit/llms/bedrock/invoke_agent/__init__.py
 create mode 100644 tests/unit/llms/bedrock/passthrough/__init__.py
 create mode 100644 tests/unit/llms/bedrock/passthrough/guardrail_translation/__init__.py
 create mode 100644 tests/unit/llms/bedrock/realtime/__init__.py
 create mode 100644 tests/unit/llms/bedrock/rerank/__init__.py
 create mode 100644 tests/unit/llms/bedrock/vector_stores/__init__.py
 create mode 100644 tests/unit/llms/bedrock_mantle/__init__.py
 create mode 100644 tests/unit/llms/bedrock_mantle/passthrough/__init__.py
 create mode 100644 tests/unit/llms/black_forest_labs/__init__.py
 create mode 100644 tests/unit/llms/black_forest_labs/image_edit/__init__.py
 create mode 100644 tests/unit/llms/black_forest_labs/image_generation/__init__.py
 create mode 100644 tests/unit/llms/bytez/__init__.py
 create mode 100644 tests/unit/llms/bytez/chat/__init__.py
 create mode 100644 tests/unit/llms/cerebras/__init__.py
 create mode 100644 tests/unit/llms/chat/__init__.py
 create mode 100644 tests/unit/llms/chatgpt/__init__.py
 create mode 100644 tests/unit/llms/chatgpt/chat/__init__.py
 create mode 100644 tests/unit/llms/chatgpt/responses/__init__.py
 create mode 100644 tests/unit/llms/cloudflare/__init__.py
 create mode 100644 tests/unit/llms/cohere/__init__.py
 create mode 100644 tests/unit/llms/cohere/chat/__init__.py
 create mode 100644 tests/unit/llms/cohere/embed/__init__.py
 create mode 100644 tests/unit/llms/cohere/ocr/__init__.py
 create mode 100644 tests/unit/llms/cohere/rerank/__init__.py
 create mode 100644 tests/unit/llms/crusoe/__init__.py
 create mode 100644 tests/unit/llms/databricks/__init__.py
 create mode 100644 tests/unit/llms/databricks/chat/__init__.py
 create mode 100644 tests/unit/llms/databricks/responses/__init__.py
 create mode 100644 tests/unit/llms/datarobot/__init__.py
 create mode 100644 tests/unit/llms/datarobot/chat/__init__.py
 create mode 100644 tests/unit/llms/deepseek/__init__.py
 create mode 100644 tests/unit/llms/deepseek/chat/__init__.py
 create mode 100644 tests/unit/llms/deepseek/messages/__init__.py
 create mode 100644 tests/unit/llms/docker_model_runner/__init__.py
 create mode 100644 tests/unit/llms/elevenlabs/__init__.py
 create mode 100644 tests/unit/llms/fastcrw/__init__.py
 create mode 100644 tests/unit/llms/fastcrw/search/__init__.py
 create mode 100644 tests/unit/llms/fireworks_ai/__init__.py
 create mode 100644 tests/unit/llms/fireworks_ai/chat/__init__.py
 create mode 100644 tests/unit/llms/fireworks_ai/rerank/__init__.py
 create mode 100644 tests/unit/llms/fireworks_ai/responses/__init__.py
 create mode 100644 tests/unit/llms/gemini/__init__.py
 create mode 100644 tests/unit/llms/gemini/audio_transcription/__init__.py
 create mode 100644 tests/unit/llms/gemini/files/__init__.py
 create mode 100644 tests/unit/llms/gemini/google_genai/__init__.py
 create mode 100644 tests/unit/llms/gemini/google_genai/guardrail_translation/__init__.py
 create mode 100644 tests/unit/llms/gemini/image_edit/__init__.py
 create mode 100644 tests/unit/llms/gemini/realtime/__init__.py
 create mode 100644 tests/unit/llms/gemini/videos/__init__.py
 create mode 100644 tests/unit/llms/gigachat/__init__.py
 create mode 100644 tests/unit/llms/gigachat/chat/__init__.py
 create mode 100644 tests/unit/llms/gigachat/embedding/__init__.py
 create mode 100644 tests/unit/llms/gigachat/passthrough/__init__.py
 create mode 100644 tests/unit/llms/github_copilot/__init__.py
 create mode 100644 tests/unit/llms/github_copilot/embedding/__init__.py
 create mode 100644 tests/unit/llms/github_copilot/messages/__init__.py
 create mode 100644 tests/unit/llms/github_copilot/responses/__init__.py
 create mode 100644 tests/unit/llms/gradient_ai/__init__.py
 create mode 100644 tests/unit/llms/gradient_ai/chat/__init__.py
 create mode 100644 tests/unit/llms/groq/__init__.py
 create mode 100644 tests/unit/llms/groq/chat/__init__.py
 create mode 100644 tests/unit/llms/hosted_vllm/__init__.py
 create mode 100644 tests/unit/llms/hosted_vllm/chat/__init__.py
 create mode 100644 tests/unit/llms/hosted_vllm/embedding/__init__.py
 create mode 100644 tests/unit/llms/hosted_vllm/image_edit/__init__.py
 create mode 100644 tests/unit/llms/hosted_vllm/responses/__init__.py
 create mode 100644 tests/unit/llms/hosted_vllm/videos/__init__.py
 create mode 100644 tests/unit/llms/huggingface/__init__.py
 create mode 100644 tests/unit/llms/huggingface/rerank/__init__.py
 create mode 100644 tests/unit/llms/inception/__init__.py
 create mode 100644 tests/unit/llms/jina_ai/__init__.py
 create mode 100644 tests/unit/llms/jina_ai/embedding/__init__.py
 create mode 100644 tests/unit/llms/langflow/__init__.py
 create mode 100644 tests/unit/llms/langflow/chat/__init__.py
 create mode 100644 tests/unit/llms/litellm_proxy/__init__.py
 create mode 100644 tests/unit/llms/litellm_proxy/chat/__init__.py
 create mode 100644 tests/unit/llms/litellm_proxy/skills/__init__.py
 create mode 100644 tests/unit/llms/llamafile/__init__.py
 create mode 100644 tests/unit/llms/llamafile/chat/__init__.py
 create mode 100644 tests/unit/llms/meta/__init__.py
 create mode 100644 tests/unit/llms/meta/realtime/__init__.py
 create mode 100644 tests/unit/llms/meta_llama/__init__.py
 create mode 100644 tests/unit/llms/mistral/audio_speech/__init__.py
 create mode 100644 tests/unit/llms/modelscope/__init__.py
 create mode 100644 tests/unit/llms/modelscope/image_generation/__init__.py
 create mode 100644 tests/unit/llms/mongodb/__init__.py
 create mode 100644 tests/unit/llms/mongodb/vector_stores/__init__.py
 create mode 100644 tests/unit/llms/moonshot/__init__.py
 create mode 100644 tests/unit/llms/neosantara/__init__.py
 create mode 100644 tests/unit/llms/nimble/__init__.py
 create mode 100644 tests/unit/llms/nimble/search/__init__.py
 create mode 100644 tests/unit/llms/novita/__init__.py
 create mode 100644 tests/unit/llms/novita/chat/__init__.py
 create mode 100644 tests/unit/llms/nscale/__init__.py
 create mode 100644 tests/unit/llms/nscale/chat/__init__.py
 create mode 100644 tests/unit/llms/nvidia_nim/__init__.py
 create mode 100644 tests/unit/llms/nvidia_nim/passthrough/__init__.py
 create mode 100644 tests/unit/llms/nvidia_nim/rerank/__init__.py
 create mode 100644 tests/unit/llms/nvidia_riva/__init__.py
 create mode 100644 tests/unit/llms/nvidia_riva/audio_transcription/__init__.py
 create mode 100644 tests/unit/llms/oci/__init__.py
 create mode 100644 tests/unit/llms/oci/chat/__init__.py
 create mode 100644 tests/unit/llms/oci/embed/__init__.py
 create mode 100644 tests/unit/llms/ocr/__init__.py
 create mode 100644 tests/unit/llms/ocr/guardrail_translation/__init__.py
 create mode 100644 tests/unit/llms/oobabooga/__init__.py
 create mode 100644 tests/unit/llms/oobabooga/chat/__init__.py
 create mode 100644 tests/unit/llms/openai/__init__.py
 create mode 100644 tests/unit/llms/openai/chat/__init__.py
 create mode 100644 tests/unit/llms/openai/chat/guardrail_translation/__init__.py
 create mode 100644 tests/unit/llms/openai/completion/__init__.py
 create mode 100644 tests/unit/test_package_layout.py

diff --git a/tests/unit/__init__.py b/tests/unit/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/integrations/__init__.py b/tests/unit/integrations/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/integrations/levo/__init__.py b/tests/unit/integrations/levo/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/integrations/litellm_agent/__init__.py b/tests/unit/integrations/litellm_agent/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/integrations/mavvrik_focus/__init__.py b/tests/unit/integrations/mavvrik_focus/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/integrations/opik/__init__.py b/tests/unit/integrations/opik/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/integrations/pointfive/__init__.py b/tests/unit/integrations/pointfive/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/integrations/vector_store_integrations/__init__.py b/tests/unit/integrations/vector_store_integrations/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/litellm_core_utils/__init__.py b/tests/unit/litellm_core_utils/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/litellm_core_utils/audio_utils/__init__.py b/tests/unit/litellm_core_utils/audio_utils/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/litellm_core_utils/llm_response_utils/__init__.py b/tests/unit/litellm_core_utils/llm_response_utils/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/__init__.py b/tests/unit/llms/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/a2a/__init__.py b/tests/unit/llms/a2a/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/a2a/chat/__init__.py b/tests/unit/llms/a2a/chat/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/a2a/chat/guardrail_translation/__init__.py b/tests/unit/llms/a2a/chat/guardrail_translation/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/anthropic/__init__.py b/tests/unit/llms/anthropic/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/anthropic/batches/__init__.py b/tests/unit/llms/anthropic/batches/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/base_llm/__init__.py b/tests/unit/llms/base_llm/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/base_llm/batches/__init__.py b/tests/unit/llms/base_llm/batches/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/base_llm/realtime/__init__.py b/tests/unit/llms/base_llm/realtime/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/baseten/__init__.py b/tests/unit/llms/baseten/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/baseten/chat/__init__.py b/tests/unit/llms/baseten/chat/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/bedrock/__init__.py b/tests/unit/llms/bedrock/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/bedrock/chat/__init__.py b/tests/unit/llms/bedrock/chat/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/bedrock/chat/agentcore/__init__.py b/tests/unit/llms/bedrock/chat/agentcore/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/bedrock/chat/invoke_transformations/__init__.py b/tests/unit/llms/bedrock/chat/invoke_transformations/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/bedrock/chat/mantle/__init__.py b/tests/unit/llms/bedrock/chat/mantle/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/bedrock/count_tokens/__init__.py b/tests/unit/llms/bedrock/count_tokens/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/bedrock/files/__init__.py b/tests/unit/llms/bedrock/files/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/bedrock/image/__init__.py b/tests/unit/llms/bedrock/image/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/bedrock/image_edit/__init__.py b/tests/unit/llms/bedrock/image_edit/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/bedrock/invoke_agent/__init__.py b/tests/unit/llms/bedrock/invoke_agent/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/bedrock/passthrough/__init__.py b/tests/unit/llms/bedrock/passthrough/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/bedrock/passthrough/guardrail_translation/__init__.py b/tests/unit/llms/bedrock/passthrough/guardrail_translation/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/bedrock/realtime/__init__.py b/tests/unit/llms/bedrock/realtime/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/bedrock/rerank/__init__.py b/tests/unit/llms/bedrock/rerank/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/bedrock/vector_stores/__init__.py b/tests/unit/llms/bedrock/vector_stores/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/bedrock_mantle/__init__.py b/tests/unit/llms/bedrock_mantle/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/bedrock_mantle/passthrough/__init__.py b/tests/unit/llms/bedrock_mantle/passthrough/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/black_forest_labs/__init__.py b/tests/unit/llms/black_forest_labs/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/black_forest_labs/image_edit/__init__.py b/tests/unit/llms/black_forest_labs/image_edit/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/black_forest_labs/image_generation/__init__.py b/tests/unit/llms/black_forest_labs/image_generation/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/bytez/__init__.py b/tests/unit/llms/bytez/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/bytez/chat/__init__.py b/tests/unit/llms/bytez/chat/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/cerebras/__init__.py b/tests/unit/llms/cerebras/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/chat/__init__.py b/tests/unit/llms/chat/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/chatgpt/__init__.py b/tests/unit/llms/chatgpt/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/chatgpt/chat/__init__.py b/tests/unit/llms/chatgpt/chat/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/chatgpt/responses/__init__.py b/tests/unit/llms/chatgpt/responses/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/cloudflare/__init__.py b/tests/unit/llms/cloudflare/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/cohere/__init__.py b/tests/unit/llms/cohere/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/cohere/chat/__init__.py b/tests/unit/llms/cohere/chat/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/cohere/embed/__init__.py b/tests/unit/llms/cohere/embed/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/cohere/ocr/__init__.py b/tests/unit/llms/cohere/ocr/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/cohere/rerank/__init__.py b/tests/unit/llms/cohere/rerank/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/crusoe/__init__.py b/tests/unit/llms/crusoe/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/databricks/__init__.py b/tests/unit/llms/databricks/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/databricks/chat/__init__.py b/tests/unit/llms/databricks/chat/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/databricks/responses/__init__.py b/tests/unit/llms/databricks/responses/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/datarobot/__init__.py b/tests/unit/llms/datarobot/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/datarobot/chat/__init__.py b/tests/unit/llms/datarobot/chat/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/deepseek/__init__.py b/tests/unit/llms/deepseek/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/deepseek/chat/__init__.py b/tests/unit/llms/deepseek/chat/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/deepseek/messages/__init__.py b/tests/unit/llms/deepseek/messages/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/docker_model_runner/__init__.py b/tests/unit/llms/docker_model_runner/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/elevenlabs/__init__.py b/tests/unit/llms/elevenlabs/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/fastcrw/__init__.py b/tests/unit/llms/fastcrw/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/fastcrw/search/__init__.py b/tests/unit/llms/fastcrw/search/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/fireworks_ai/__init__.py b/tests/unit/llms/fireworks_ai/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/fireworks_ai/chat/__init__.py b/tests/unit/llms/fireworks_ai/chat/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/fireworks_ai/rerank/__init__.py b/tests/unit/llms/fireworks_ai/rerank/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/fireworks_ai/responses/__init__.py b/tests/unit/llms/fireworks_ai/responses/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/gemini/__init__.py b/tests/unit/llms/gemini/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/gemini/audio_transcription/__init__.py b/tests/unit/llms/gemini/audio_transcription/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/gemini/files/__init__.py b/tests/unit/llms/gemini/files/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/gemini/google_genai/__init__.py b/tests/unit/llms/gemini/google_genai/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/gemini/google_genai/guardrail_translation/__init__.py b/tests/unit/llms/gemini/google_genai/guardrail_translation/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/gemini/image_edit/__init__.py b/tests/unit/llms/gemini/image_edit/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/gemini/realtime/__init__.py b/tests/unit/llms/gemini/realtime/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/gemini/videos/__init__.py b/tests/unit/llms/gemini/videos/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/gigachat/__init__.py b/tests/unit/llms/gigachat/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/gigachat/chat/__init__.py b/tests/unit/llms/gigachat/chat/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/gigachat/embedding/__init__.py b/tests/unit/llms/gigachat/embedding/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/gigachat/passthrough/__init__.py b/tests/unit/llms/gigachat/passthrough/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/github_copilot/__init__.py b/tests/unit/llms/github_copilot/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/github_copilot/embedding/__init__.py b/tests/unit/llms/github_copilot/embedding/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/github_copilot/messages/__init__.py b/tests/unit/llms/github_copilot/messages/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/github_copilot/responses/__init__.py b/tests/unit/llms/github_copilot/responses/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/gradient_ai/__init__.py b/tests/unit/llms/gradient_ai/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/gradient_ai/chat/__init__.py b/tests/unit/llms/gradient_ai/chat/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/groq/__init__.py b/tests/unit/llms/groq/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/groq/chat/__init__.py b/tests/unit/llms/groq/chat/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/hosted_vllm/__init__.py b/tests/unit/llms/hosted_vllm/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/hosted_vllm/chat/__init__.py b/tests/unit/llms/hosted_vllm/chat/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/hosted_vllm/embedding/__init__.py b/tests/unit/llms/hosted_vllm/embedding/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/hosted_vllm/image_edit/__init__.py b/tests/unit/llms/hosted_vllm/image_edit/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/hosted_vllm/responses/__init__.py b/tests/unit/llms/hosted_vllm/responses/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/hosted_vllm/videos/__init__.py b/tests/unit/llms/hosted_vllm/videos/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/huggingface/__init__.py b/tests/unit/llms/huggingface/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/huggingface/rerank/__init__.py b/tests/unit/llms/huggingface/rerank/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/inception/__init__.py b/tests/unit/llms/inception/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/jina_ai/__init__.py b/tests/unit/llms/jina_ai/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/jina_ai/embedding/__init__.py b/tests/unit/llms/jina_ai/embedding/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/langflow/__init__.py b/tests/unit/llms/langflow/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/langflow/chat/__init__.py b/tests/unit/llms/langflow/chat/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/litellm_proxy/__init__.py b/tests/unit/llms/litellm_proxy/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/litellm_proxy/chat/__init__.py b/tests/unit/llms/litellm_proxy/chat/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/litellm_proxy/skills/__init__.py b/tests/unit/llms/litellm_proxy/skills/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/llamafile/__init__.py b/tests/unit/llms/llamafile/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/llamafile/chat/__init__.py b/tests/unit/llms/llamafile/chat/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/meta/__init__.py b/tests/unit/llms/meta/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/meta/realtime/__init__.py b/tests/unit/llms/meta/realtime/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/meta_llama/__init__.py b/tests/unit/llms/meta_llama/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/mistral/audio_speech/__init__.py b/tests/unit/llms/mistral/audio_speech/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/modelscope/__init__.py b/tests/unit/llms/modelscope/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/modelscope/image_generation/__init__.py b/tests/unit/llms/modelscope/image_generation/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/mongodb/__init__.py b/tests/unit/llms/mongodb/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/mongodb/vector_stores/__init__.py b/tests/unit/llms/mongodb/vector_stores/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/moonshot/__init__.py b/tests/unit/llms/moonshot/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/neosantara/__init__.py b/tests/unit/llms/neosantara/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/nimble/__init__.py b/tests/unit/llms/nimble/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/nimble/search/__init__.py b/tests/unit/llms/nimble/search/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/novita/__init__.py b/tests/unit/llms/novita/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/novita/chat/__init__.py b/tests/unit/llms/novita/chat/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/nscale/__init__.py b/tests/unit/llms/nscale/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/nscale/chat/__init__.py b/tests/unit/llms/nscale/chat/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/nvidia_nim/__init__.py b/tests/unit/llms/nvidia_nim/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/nvidia_nim/passthrough/__init__.py b/tests/unit/llms/nvidia_nim/passthrough/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/nvidia_nim/rerank/__init__.py b/tests/unit/llms/nvidia_nim/rerank/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/nvidia_riva/__init__.py b/tests/unit/llms/nvidia_riva/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/nvidia_riva/audio_transcription/__init__.py b/tests/unit/llms/nvidia_riva/audio_transcription/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/oci/__init__.py b/tests/unit/llms/oci/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/oci/chat/__init__.py b/tests/unit/llms/oci/chat/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/oci/embed/__init__.py b/tests/unit/llms/oci/embed/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/ocr/__init__.py b/tests/unit/llms/ocr/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/ocr/guardrail_translation/__init__.py b/tests/unit/llms/ocr/guardrail_translation/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/oobabooga/__init__.py b/tests/unit/llms/oobabooga/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/oobabooga/chat/__init__.py b/tests/unit/llms/oobabooga/chat/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/openai/__init__.py b/tests/unit/llms/openai/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/openai/chat/__init__.py b/tests/unit/llms/openai/chat/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/openai/chat/guardrail_translation/__init__.py b/tests/unit/llms/openai/chat/guardrail_translation/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/llms/openai/completion/__init__.py b/tests/unit/llms/openai/completion/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/unit/test_package_layout.py b/tests/unit/test_package_layout.py
new file mode 100644
index 00000000000..4ea68fc06ba
--- /dev/null
+++ b/tests/unit/test_package_layout.py
@@ -0,0 +1,12 @@
+import os
+
+TESTS_UNIT_DIR = os.path.dirname(os.path.abspath(__file__))
+
+
+def test_every_directory_under_tests_unit_is_a_package():
+    missing = []
+    for root, dirs, _files in os.walk(TESTS_UNIT_DIR):
+        dirs[:] = [d for d in dirs if d != "__pycache__"]
+        if not os.path.isfile(os.path.join(root, "__init__.py")):
+            missing.append(os.path.relpath(root, TESTS_UNIT_DIR))
+    assert missing == []

From 73a35abeb303dff68f232719bac0893cddf69366 Mon Sep 17 00:00:00 2001
From: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Date: Sun, 20 Sep 2026 11:57:53 +0000
Subject: [PATCH 13/13] refactor(types): keep object permission parsing as it
 was

Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
---
 .../management_helpers/object_permission_utils.py   | 13 ++++++-------
 1 file changed, 6 insertions(+), 7 deletions(-)

diff --git a/litellm/proxy/management_helpers/object_permission_utils.py b/litellm/proxy/management_helpers/object_permission_utils.py
index 61e432daa16..daab38d3662 100644
--- a/litellm/proxy/management_helpers/object_permission_utils.py
+++ b/litellm/proxy/management_helpers/object_permission_utils.py
@@ -8,7 +8,7 @@ from collections.abc import Mapping, Sequence
 from collections.abc import Set as AbstractSet
 from dataclasses import dataclass
 from types import MappingProxyType
-from typing import TYPE_CHECKING, Final, Optional
+from typing import TYPE_CHECKING, Any, Final, Optional
 
 from fastapi import HTTPException, status
 from pydantic import TypeAdapter
@@ -156,13 +156,12 @@ async def handle_update_object_permission_common(
     if prisma_client is None:
         raise ValueError("Prisma client not found")
 
-    raw_object_permission: Final[dict | str | None] = data_json.pop("object_permission", None)
-    if raw_object_permission is None:
+    new_object_permission: dict | str | None = data_json.pop("object_permission", None)
+    if new_object_permission is None:
         return None
 
-    new_object_permission: Final[object] = (
-        json.loads(raw_object_permission) if isinstance(raw_object_permission, str) else raw_object_permission
-    )
+    if isinstance(new_object_permission, str):
+        new_object_permission = json.loads(new_object_permission)
 
     upsert: Final = await prepare_object_permission_upsert(
         new_object_permission=new_object_permission if isinstance(new_object_permission, dict) else {},
@@ -231,7 +230,7 @@ def _dedupe_preserving_order(values: list[str]) -> list[str]:
     return result
 
 
-def _mcp_server_identifier_matches(server: object, identifier: str) -> bool:
+def _mcp_server_identifier_matches(server: Any, identifier: str) -> bool:
     return identifier in {
         getattr(server, "server_id", None),
         getattr(server, "alias", None),