diff --git a/backend/open_webui/routers/ollama.py b/backend/open_webui/routers/ollama.py
index 737ed7153..b3f0e7ec9 100644
--- a/backend/open_webui/routers/ollama.py
+++ b/backend/open_webui/routers/ollama.py
@@ -38,7 +38,7 @@ from fastapi import (
 from fastapi.middleware.cors import CORSMiddleware
 from fastapi.responses import StreamingResponse
 from pydantic import BaseModel, ConfigDict, validator
-from starlette.background import BackgroundTask
+
 from sqlalchemy.orm import Session
 
 from open_webui.internal.db import get_session
@@ -49,6 +49,8 @@ from open_webui.models.access_grants import AccessGrants
 from open_webui.models.groups import Groups
 from open_webui.utils.misc import (
     calculate_sha256,
+    cleanup_response,
+    stream_wrapper,
 )
 from open_webui.utils.payload import (
     apply_model_params_to_body_ollama,
@@ -103,14 +105,6 @@ async def send_get_request(url, key=None, user: UserModel = None):
         return None
 
 
-async def cleanup_response(
-    response: Optional[aiohttp.ClientResponse],
-    session: Optional[aiohttp.ClientSession],
-):
-    if response:
-        response.close()
-    if session:
-        await session.close()
 
 
 async def send_post_request(
@@ -124,6 +118,7 @@ async def send_post_request(
 ):
 
     r = None
+    streaming = False
     try:
         session = aiohttp.ClientSession(
             trust_env=True, timeout=aiohttp.ClientTimeout(total=AIOHTTP_CLIENT_TIMEOUT)
@@ -168,13 +163,11 @@ async def send_post_request(
             if content_type:
                 response_headers["Content-Type"] = content_type
 
+            streaming = True
             return StreamingResponse(
-                r.content,
+                stream_wrapper(r, session),
                 status_code=r.status,
                 headers=response_headers,
-                background=BackgroundTask(
-                    cleanup_response, response=r, session=session
-                ),
             )
         else:
             res = await r.json()
@@ -190,7 +183,7 @@ async def send_post_request(
             detail=detail if e else "Open WebUI: Server Connection Error",
         )
     finally:
-        if not stream:
+        if not streaming:
             await cleanup_response(r, session)
 
 
diff --git a/backend/open_webui/routers/openai.py b/backend/open_webui/routers/openai.py
index 38fd7feed..d5f3aa88b 100644
--- a/backend/open_webui/routers/openai.py
+++ b/backend/open_webui/routers/openai.py
@@ -18,7 +18,7 @@ from fastapi.responses import (
     PlainTextResponse,
 )
 from pydantic import BaseModel
-from starlette.background import BackgroundTask
+
 from sqlalchemy.orm import Session
 
 from open_webui.internal.db import get_session
@@ -48,8 +48,10 @@ from open_webui.utils.payload import (
     apply_system_prompt_to_body,
 )
 from open_webui.utils.misc import (
+    cleanup_response,
     convert_logit_bias_input_to_json,
     stream_chunks_handler,
+    stream_wrapper,
 )
 
 from open_webui.utils.auth import get_admin_user, get_verified_user
@@ -88,14 +90,6 @@ async def send_get_request(url, key=None, user: UserModel = None):
         return None
 
 
-async def cleanup_response(
-    response: Optional[aiohttp.ClientResponse],
-    session: Optional[aiohttp.ClientSession],
-):
-    if response:
-        response.close()
-    if session:
-        await session.close()
 
 
 def openai_reasoning_model_handler(payload):
@@ -1104,12 +1098,9 @@ async def generate_chat_completion(
         if "text/event-stream" in r.headers.get("Content-Type", ""):
             streaming = True
             return StreamingResponse(
-                stream_chunks_handler(r.content),
+                stream_wrapper(r, session, stream_chunks_handler),
                 status_code=r.status,
                 headers=dict(r.headers),
-                background=BackgroundTask(
-                    cleanup_response, response=r, session=session
-                ),
             )
         else:
             try:
@@ -1190,12 +1181,9 @@ async def embeddings(request: Request, form_data: dict, user):
         if "text/event-stream" in r.headers.get("Content-Type", ""):
             streaming = True
             return StreamingResponse(
-                r.content,
+                stream_wrapper(r, session),
                 status_code=r.status,
                 headers=dict(r.headers),
-                background=BackgroundTask(
-                    cleanup_response, response=r, session=session
-                ),
             )
         else:
             try:
@@ -1282,12 +1270,9 @@ async def proxy(path: str, request: Request, user=Depends(get_verified_user)):
         if "text/event-stream" in r.headers.get("Content-Type", ""):
             streaming = True
             return StreamingResponse(
-                r.content,
+                stream_wrapper(r, session),
                 status_code=r.status,
                 headers=dict(r.headers),
-                background=BackgroundTask(
-                    cleanup_response, response=r, session=session
-                ),
             )
         else:
             try:
diff --git a/backend/open_webui/utils/misc.py b/backend/open_webui/utils/misc.py
index 756dd9161..8e95e9bfc 100644
--- a/backend/open_webui/utils/misc.py
+++ b/backend/open_webui/utils/misc.py
@@ -782,6 +782,30 @@ def extract_urls(text: str) -> list[str]:
     return url_pattern.findall(text)
 
 
+
+async def cleanup_response(
+    response: Optional[aiohttp.ClientResponse],
+    session: Optional[aiohttp.ClientSession],
+):
+    if response:
+        response.close()
+    if session:
+        await session.close()
+
+
+async def stream_wrapper(response, session, content_handler=None):
+    """
+    Wrap a stream to ensure cleanup happens even if streaming is interrupted.
+    This is more reliable than BackgroundTask which may not run if client disconnects.
+    """
+    try:
+        stream = content_handler(response.content) if content_handler else response.content
+        async for chunk in stream:
+            yield chunk
+    finally:
+        await cleanup_response(response, session)
+
+
 def stream_chunks_handler(stream: aiohttp.StreamReader):
     """
     Handle stream response chunks, supporting large data chunks that exceed the original 16kb limit.