diff --git a/tests/llm_translation/realtime/base_realtime_tests.py b/tests/llm_translation/realtime/base_realtime_tests.py
index 1d55f13b00..f1c4265900 100644
--- a/tests/llm_translation/realtime/base_realtime_tests.py
+++ b/tests/llm_translation/realtime/base_realtime_tests.py
@@ -79,7 +79,7 @@ class RealTimeWebSocketClient:
 
     def _is_initial_event(self, msg_type: str) -> bool:
         """Check if message type is an initial connection event"""
-        # OpenAI sends "session.created", xAI sends "conversation.created"
+        # OpenAI and xAI send "session.created"; some providers send "conversation.created"
         return msg_type in ["session.created", "conversation.created"]
 
     async def receive_text(self):
diff --git a/tests/llm_translation/realtime/test_xai_realtime.py b/tests/llm_translation/realtime/test_xai_realtime.py
index 0bb7a59bb1..86d0ebe3a3 100644
--- a/tests/llm_translation/realtime/test_xai_realtime.py
+++ b/tests/llm_translation/realtime/test_xai_realtime.py
@@ -19,8 +19,8 @@ class TestXAIRealtime(BaseRealtimeTest):
     """
     E2E tests for xAI Realtime API.
 
-    xAI's Grok Voice Agent API is OpenAI-compatible but uses:
-    - Different initial event: "conversation.created" instead of "session.created"
+    xAI's Grok Voice Agent API is OpenAI-compatible:
+    - Initial event: "session.created" (matches OpenAI)
     - Different endpoint: wss://api.x.ai/v1/realtime
     - Model: grok-4-1-fast-non-reasoning
     """
@@ -32,4 +32,4 @@ class TestXAIRealtime(BaseRealtimeTest):
         return "XAI_API_KEY"
 
     def get_initial_event_type(self) -> str:
-        return "conversation.created"
+        return "session.created"