diff --git a/vllm/transformers_utils/tokenizers/mistral.py b/vllm/transformers_utils/tokenizers/mistral.py
index 8d49c13afe301..946e4d51f7ac5 100644
--- a/vllm/transformers_utils/tokenizers/mistral.py
+++ b/vllm/transformers_utils/tokenizers/mistral.py
@@ -52,11 +52,12 @@ def __init__(self, tokenizer: PublicMistralTokenizer) -> None:
         assert isinstance(self.tokenizer,
                           (Tekkenizer, SentencePieceTokenizer)), type(
                               self.tokenizer)
-        self._is_tekken = isinstance(self.tokenizer, Tekkenizer)
 
-        if self._is_tekken:
+        if (is_tekken := isinstance(self.tokenizer, Tekkenizer)):
             # Make sure special tokens will not raise
             self.tokenizer.special_token_policy = SpecialTokenPolicy.IGNORE
+        
+        self._is_tekken = is_tekken
 
         # the following attributes are set to fit VLLM's design
         self.is_fast = True