diff --git a/vllm/v1/attention/backends/utils.py b/vllm/v1/attention/backends/utils.py
index 1b030eaf1..42459815e 100644
--- a/vllm/v1/attention/backends/utils.py
+++ b/vllm/v1/attention/backends/utils.py
@@ -528,7 +528,6 @@ def split_decodes_and_prefills(
         # requests may have a query length of 0 but since they are padding its fine
         # to treat them as decodes (ensures num_decodes matches the captured size)
         if torch.all((query_lens == query_lens[0]) | (query_lens == 0)):
-            assert num_reqs * query_lens[0] == num_tokens, "tokens not padded correctly"
             return num_reqs, 0, num_tokens, 0  # all decodes
         is_prefill = query_lens != query_lens[0]
     else: