diff --git a/litellm/model_prices_and_context_window_backup.json b/litellm/model_prices_and_context_window_backup.json
index 8e5b8251f6f..8cab321c4dd 100644
--- a/litellm/model_prices_and_context_window_backup.json
+++ b/litellm/model_prices_and_context_window_backup.json
@@ -21915,8 +21915,8 @@
         "input_cost_per_token_batches": 7.5e-06,
         "litellm_provider": "vertex_ai-anthropic_models",
         "max_input_tokens": 200000,
-        "max_output_tokens": 4096,
-        "max_tokens": 4096,
+        "max_output_tokens": 32000,
+        "max_tokens": 32000,
         "mode": "chat",
         "output_cost_per_token": 7.5e-05,
         "output_cost_per_token_batches": 3.75e-05,
@@ -21932,8 +21932,8 @@
         "input_cost_per_token_batches": 7.5e-06,
         "litellm_provider": "vertex_ai-anthropic_models",
         "max_input_tokens": 200000,
-        "max_output_tokens": 4096,
-        "max_tokens": 4096,
+        "max_output_tokens": 32000,
+        "max_tokens": 32000,
         "mode": "chat",
         "output_cost_per_token": 7.5e-05,
         "output_cost_per_token_batches": 3.75e-05,
diff --git a/litellm/proxy/management_endpoints/ui_sso.py b/litellm/proxy/management_endpoints/ui_sso.py
index 54909b9712b..541f7656d20 100644
--- a/litellm/proxy/management_endpoints/ui_sso.py
+++ b/litellm/proxy/management_endpoints/ui_sso.py
@@ -875,9 +875,9 @@ async def insert_sso_user(
         if user_defined_values.get("max_budget") is None:
             user_defined_values["max_budget"] = litellm.max_internal_user_budget
         if user_defined_values.get("budget_duration") is None:
-            user_defined_values[
-                "budget_duration"
-            ] = litellm.internal_user_budget_duration
+            user_defined_values["budget_duration"] = (
+                litellm.internal_user_budget_duration
+            )
 
     if user_defined_values["user_role"] is None:
         user_defined_values["user_role"] = LitellmUserRoles.INTERNAL_USER_VIEW_ONLY
@@ -1118,9 +1118,9 @@ class SSOAuthenticationHandler:
                 generic_authorization_endpoint
                 and "okta" in generic_authorization_endpoint
             ):
-                redirect_params[
-                    "state"
-                ] = uuid.uuid4().hex  # set state param for okta - required
+                redirect_params["state"] = (
+                    uuid.uuid4().hex
+                )  # set state param for okta - required
 
         return redirect_params
 
@@ -1699,9 +1699,9 @@ class MicrosoftSSOHandler:
 
         # if user is trying to get the raw sso response for debugging, return the raw sso response
         if return_raw_sso_response:
-            original_msft_result[
-                MicrosoftSSOHandler.GRAPH_API_RESPONSE_KEY
-            ] = user_team_ids
+            original_msft_result[MicrosoftSSOHandler.GRAPH_API_RESPONSE_KEY] = (
+                user_team_ids
+            )
             original_msft_result["app_roles"] = app_roles
             return original_msft_result or {}
 
@@ -1759,7 +1759,8 @@ class MicrosoftSSOHandler:
             decoded_token = jwt.decode(id_token, options={"verify_signature": False})
 
             # Extract app_roles claim from the token
-            roles = decoded_token.get("app_roles", [])
+            ## check for both 'roles' and 'app_roles' claims
+            roles = decoded_token.get("app_roles", []) or decoded_token.get("roles", [])
 
             if roles and isinstance(roles, list):
                 verbose_proxy_logger.debug(
@@ -1817,9 +1818,9 @@ class MicrosoftSSOHandler:
 
             # Fetch user membership from Microsoft Graph API
             all_group_ids = []
-            next_link: Optional[
-                str
-            ] = MicrosoftSSOHandler.graph_api_user_groups_endpoint
+            next_link: Optional[str] = (
+                MicrosoftSSOHandler.graph_api_user_groups_endpoint
+            )
             auth_headers = {"Authorization": f"Bearer {access_token}"}
             page_count = 0