diff --git a/litellm/litellm_core_utils/realtime_streaming.py b/litellm/litellm_core_utils/realtime_streaming.py
index a35d113fc9d..97a87ca1faf 100644
--- a/litellm/litellm_core_utils/realtime_streaming.py
+++ b/litellm/litellm_core_utils/realtime_streaming.py
@@ -5,6 +5,7 @@ from typing import TYPE_CHECKING, Any, Dict, List, Optional, Union, cast
 
 import litellm
 from litellm._logging import verbose_logger
+from litellm.litellm_core_utils.logging_worker import GLOBAL_LOGGING_WORKER
 from litellm.llms.base_llm.realtime.transformation import BaseRealtimeConfig
 from litellm.types.llms.openai import (
     OpenAIRealtimeEvents,
@@ -327,8 +328,10 @@ class RealTimeStreaming:
                     self.tool_calls
                 )
             ## ASYNC LOGGING
-            # Create an event loop for the new thread
-            asyncio.create_task(self.logging_obj.async_success_handler(self.messages))
+            # Route through the bounded logging worker (per-coroutine timeout +
+            # concurrency cap) instead of a bare create_task, so a slow callback
+            # can't leave suspended tasks pinning each call's response in memory.
+            GLOBAL_LOGGING_WORKER.ensure_initialized_and_enqueue(self.logging_obj.async_success_handler(self.messages))
             ## SYNC LOGGING
             executor.submit(self.logging_obj.success_handler(self.messages))
 
diff --git a/tests/test_litellm/litellm_core_utils/test_realtime_streaming.py b/tests/test_litellm/litellm_core_utils/test_realtime_streaming.py
index 18547f2f6a3..cad5f6d5b38 100644
--- a/tests/test_litellm/litellm_core_utils/test_realtime_streaming.py
+++ b/tests/test_litellm/litellm_core_utils/test_realtime_streaming.py
@@ -2750,3 +2750,26 @@ def test_non_bidi_setup_left_untouched_for_followup_capable_providers():
         assert streaming._maybe_inject_guardrail_auto_response_disable(msg) == msg
     finally:
         litellm.callbacks = []
+
+
+@pytest.mark.asyncio
+async def test_log_messages_routes_async_logging_through_bounded_worker():
+    """Realtime success logging must go through GLOBAL_LOGGING_WORKER (bounded
+    queue + per-coroutine timeout), not a bare asyncio.create_task. A bare task
+    has no timeout/concurrency cap, so when a logging callback is slow every
+    realtime turn leaves a suspended task pinning its response in memory -> an
+    unbounded leak. Regression for that fix."""
+    logging_obj = MagicMock()
+    streaming = RealTimeStreaming(MagicMock(), MagicMock(), logging_obj)
+    streaming.messages = [{"type": "session.created"}]
+
+    with (
+        patch("litellm.litellm_core_utils.realtime_streaming.GLOBAL_LOGGING_WORKER") as mock_worker,
+        patch("litellm.litellm_core_utils.realtime_streaming.asyncio.create_task") as mock_create_task,
+        patch("litellm.litellm_core_utils.realtime_streaming.executor.submit"),
+    ):
+        await streaming.log_messages()
+
+        mock_worker.ensure_initialized_and_enqueue.assert_called_once()
+        # the bare create_task path must no longer be used for success logging
+        mock_create_task.assert_not_called()