diff --git a/vllm/config.py b/vllm/config.py
index 7a5b866a11d9d..1396055997c45 100644
--- a/vllm/config.py
+++ b/vllm/config.py
@@ -1819,10 +1819,6 @@ class ParallelConfig:
     enable_microbatching: bool = False
     """Enable microbatching for the model executor."""
 
-    always_microbatch_if_enabled: bool = True
-    """Always microbatch if microbatching is enabled. Easier to sync between
-       dp workers."""
-
     microbatching_token_threshold: int = 4
     """The threshold for microbatching. If the number of tokens in the
     request is greater than this threshold, microbatching will be used.