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.