diff --git a/backend/open_webui/main.py b/backend/open_webui/main.py
index f19bffc019..d040371b1f 100644
--- a/backend/open_webui/main.py
+++ b/backend/open_webui/main.py
@@ -241,7 +241,7 @@ from open_webui.utils.middleware import (
     process_chat_payload,
     process_chat_response,
 )
-from open_webui.utils.misc import merge_model_params
+from open_webui.utils.misc import get_response_error_detail, merge_model_params
 from open_webui.utils.model_ids import strip_provider_model_prefix
 from open_webui.utils.models import (
     check_model_access,
@@ -1638,14 +1638,7 @@ async def chat_completion(
             # raise so the except-block below emits chat:message:error +
             # chat:tasks:cancel, unblocking the frontend.
             if isinstance(response, JSONResponse) and response.status_code >= 400:
-                try:
-                    error_body = JSONCodec.loads(response.body.decode('utf-8', 'replace'))
-                    detail = error_body.get('error', error_body) if isinstance(error_body, dict) else error_body
-                    if isinstance(detail, dict):
-                        detail = detail.get('message', detail.get('detail', str(detail)))
-                except Exception:
-                    detail = f'Provider returned HTTP {response.status_code}'
-                raise Exception(detail)
+                raise Exception(get_response_error_detail(response))
 
             ctx = await build_chat_response_context(request, form_data, user, model, metadata, tasks, events)
 
diff --git a/backend/open_webui/utils/middleware.py b/backend/open_webui/utils/middleware.py
index 4d5eeebcd7..092c2dd3b8 100644
--- a/backend/open_webui/utils/middleware.py
+++ b/backend/open_webui/utils/middleware.py
@@ -111,6 +111,7 @@ from open_webui.utils.misc import (
     get_last_user_message_item,
     get_message_list,
     get_output_text,
+    get_response_error_detail,
     get_reasoning_details,
     get_system_message,
     is_string_allowed,
@@ -5958,6 +5959,9 @@ async def streaming_chat_response_handler(response, ctx):
                             await stream_body_handler(res, new_form_data)
                             output[:0] = prior_output
                             prior_output = []
+                        elif getattr(res, 'status_code', 200) >= 400:
+                            await emit_message_error(get_message_error_content(get_response_error_detail(res)))
+                            break
                         else:
                             break
                     except Exception as e:
@@ -6142,6 +6146,9 @@ async def streaming_chat_response_handler(response, ctx):
 
                             if isinstance(res, StreamingResponse):
                                 await stream_body_handler(res, new_form_data)
+                            elif getattr(res, 'status_code', 200) >= 400:
+                                await emit_message_error(get_message_error_content(get_response_error_detail(res)))
+                                break
                             else:
                                 break
                         except Exception as e:
diff --git a/backend/open_webui/utils/misc.py b/backend/open_webui/utils/misc.py
index 2761adc191..beb91d2b04 100644
--- a/backend/open_webui/utils/misc.py
+++ b/backend/open_webui/utils/misc.py
@@ -40,6 +40,31 @@ def merge_model_params(base: dict, override: dict) -> dict:
     return params
 
 
+def get_response_error_detail(response: object) -> str:
+    status_code = getattr(response, 'status_code', None)
+    fallback = f'Provider returned HTTP {status_code}' if status_code else 'Provider returned an error'
+
+    try:
+        body = response.body
+        if not isinstance(body, str):
+            body = body.decode('utf-8', 'replace')
+        detail = JSONCodec.loads(body)
+    except Exception:
+        return fallback
+
+    while isinstance(detail, dict):
+        next_detail = None
+        for key in ('error', 'message', 'detail'):
+            if key in detail:
+                next_detail = detail[key]
+                break
+        if next_detail is None:
+            return str(detail)
+        detail = next_detail
+
+    return detail if isinstance(detail, str) else str(detail)
+
+
 def _strip_filter_entry(entry):
     # Compose list-form env syntax passes surrounding quotes through verbatim
     return (entry or '').strip().strip('"\'').strip()