diff --git a/enterprise/litellm_enterprise/proxy/hooks/managed_files.py b/enterprise/litellm_enterprise/proxy/hooks/managed_files.py
index ca73a0574da..4841fad2ec9 100644
--- a/enterprise/litellm_enterprise/proxy/hooks/managed_files.py
+++ b/enterprise/litellm_enterprise/proxy/hooks/managed_files.py
@@ -67,7 +67,6 @@ from litellm.types.llms.openai import (  # pyright: ignore[reportAttributeAccess
     CreateFileRequest,
     FileObject,
     OpenAIFileObject,
-    OpenAIFilesPurpose,
     ResponsesAPIResponse,
 )
 from litellm.types.utils import (
@@ -1386,7 +1385,10 @@ class _PROXY_LiteLLMManagedFiles(CustomLogger, BaseFileEndpoints):
         Pagination is keyset based on ``unified_file_id`` so a key that owns
         every file on the proxy still reads one bounded page at a time.
         ``purpose`` is applied after parsing because the managed file table
-        keeps it inside the ``file_object`` blob instead of a column.
+        keeps it inside the ``file_object`` blob instead of a column, so a
+        narrowed page can hold fewer files than ``limit``. ``last_id`` then
+        falls back to the last row the page read, which keeps the cursor
+        usable even when every file on the page was filtered out.
         """
         validate_file_list_limit(limit)
         if limit == 0:
@@ -1416,14 +1418,19 @@ class _PROXY_LiteLLMManagedFiles(CustomLogger, BaseFileEndpoints):
             **cursor_args,
         )
         has_more: Final = len(rows) > page_size
+        page_rows: Final = rows[:page_size]
 
         files: Final = [
             parsed_file_object.model_copy(update={"id": row.unified_file_id})
-            for row in rows[:page_size]
+            for row in page_rows
             if (parsed_file_object := _parse_managed_file_object(row.file_object, row.unified_file_id)) is not None
             and (purpose is None or parsed_file_object.purpose == purpose)
         ]
-        return build_list_page(files, has_more=has_more)
+        return build_list_page(
+            files,
+            has_more=has_more,
+            next_cursor_id=page_rows[-1].unified_file_id if page_rows else None,
+        )
 
     def _is_batch_polling_enabled(self) -> bool:
         """
diff --git a/litellm/llms/base_llm/managed_resources/isolation.py b/litellm/llms/base_llm/managed_resources/isolation.py
index e1b204214d7..f1a54943de0 100644
--- a/litellm/llms/base_llm/managed_resources/isolation.py
+++ b/litellm/llms/base_llm/managed_resources/isolation.py
@@ -19,15 +19,23 @@ from litellm.proxy._types import (
 )
 
 
-def build_list_page(items: list[Any], has_more: bool = False) -> dict[str, Any]:
+def build_list_page(
+    items: list[Any],
+    has_more: bool = False,
+    next_cursor_id: str | None = None,
+) -> dict[str, Any]:
     """Build the OpenAI-style paginated list response shape used by managed
     file/batch/vector-store listings. ``first_id`` and ``last_id`` are
-    sourced from each item's ``.id`` attribute."""
+    sourced from each item's ``.id`` attribute.
+
+    A listing that filters rows out after reading them can pass
+    ``next_cursor_id`` so an empty page still carries the cursor the caller
+    needs to reach the rows behind it."""
     return {
         "object": "list",
         "data": items,
         "first_id": items[0].id if items else None,
-        "last_id": items[-1].id if items else None,
+        "last_id": items[-1].id if items else next_cursor_id,
         "has_more": has_more,
     }
 
diff --git a/tests/test_litellm/enterprise/proxy/test_managed_files_hook.py b/tests/test_litellm/enterprise/proxy/test_managed_files_hook.py
index 49a1119c7a0..68ded79199d 100644
--- a/tests/test_litellm/enterprise/proxy/test_managed_files_hook.py
+++ b/tests/test_litellm/enterprise/proxy/test_managed_files_hook.py
@@ -334,6 +334,40 @@ async def test_afile_list_filters_by_purpose():
     assert [file.id for file in response["data"]] == ["unified-batch"]
 
 
+@pytest.mark.asyncio
+async def test_afile_list_keeps_a_usable_cursor_when_a_page_filters_everything_out():
+    managed_files, _ = _make_managed_files_over_rows(
+        [
+            _make_managed_file_row("unified-0"),
+            _make_managed_file_row("unified-1"),
+            _make_managed_file_row("unified-2", purpose="batch"),
+        ]
+    )
+    user_api_key_dict = _make_user_api_key_dict()
+
+    first_page = await managed_files.afile_list(
+        purpose="batch",
+        litellm_parent_otel_span=None,
+        user_api_key_dict=user_api_key_dict,
+        limit=2,
+    )
+
+    assert first_page["data"] == []
+    assert first_page["has_more"] is True
+    assert first_page["last_id"] == "unified-1"
+
+    second_page = await managed_files.afile_list(
+        purpose="batch",
+        litellm_parent_otel_span=None,
+        user_api_key_dict=user_api_key_dict,
+        limit=2,
+        after=first_page["last_id"],
+    )
+
+    assert [file.id for file in second_page["data"]] == ["unified-2"]
+    assert second_page["has_more"] is False
+
+
 @pytest.mark.asyncio
 async def test_afile_list_honors_limit_and_reports_more_pages():
     managed_files, table = _make_managed_files_over_rows(