diff --git a/test/components/generators/chat/test_hugging_face_api.py b/test/components/generators/chat/test_hugging_face_api.py
index abe3f2e26..dc0f8c706 100644
--- a/test/components/generators/chat/test_hugging_face_api.py
+++ b/test/components/generators/chat/test_hugging_face_api.py
@@ -822,6 +822,7 @@ class TestHuggingFaceAPIChatGenerator:
         reason="Export an env var called HF_API_TOKEN containing the Hugging Face token to run this test.",
     )
     @pytest.mark.flaky(reruns=3, reruns_delay=10)
+    @pytest.mark.asyncio
     async def test_live_run_async_serverless(self):
         generator = HuggingFaceAPIChatGenerator(
             api_type=HFGenerationAPIType.SERVERLESS_INFERENCE_API,
@@ -829,17 +830,18 @@ class TestHuggingFaceAPIChatGenerator:
             generation_kwargs={"max_tokens": 20},
         )
 
-        # No need for instruction tokens here since we use the chat_completion endpoint which handles the chat
-        # templating for us.
         messages = [
             ChatMessage.from_user("What is the capital of France? Be concise only provide the capital, nothing else.")
         ]
-        response = await generator.run_async(messages=messages)
+        try:
+            response = await generator.run_async(messages=messages)
 
-        assert "replies" in response
-        assert isinstance(response["replies"], list)
-        assert len(response["replies"]) > 0
-        assert [isinstance(reply, ChatMessage) for reply in response["replies"]]
-        assert "usage" in response["replies"][0].meta
-        assert "prompt_tokens" in response["replies"][0].meta["usage"]
-        assert "completion_tokens" in response["replies"][0].meta["usage"]
+            assert "replies" in response
+            assert isinstance(response["replies"], list)
+            assert len(response["replies"]) > 0
+            assert [isinstance(reply, ChatMessage) for reply in response["replies"]]
+            assert "usage" in response["replies"][0].meta
+            assert "prompt_tokens" in response["replies"][0].meta["usage"]
+            assert "completion_tokens" in response["replies"][0].meta["usage"]
+        finally:
+            await generator._async_client.close()