diff --git a/vllm_ascend/worker/model_runner_v1.py b/vllm_ascend/worker/model_runner_v1.py
index d3a29852bd6..d52be032707 100644
--- a/vllm_ascend/worker/model_runner_v1.py
+++ b/vllm_ascend/worker/model_runner_v1.py
@@ -219,7 +219,7 @@ def __init__(self, vllm_config: VllmConfig, device: torch.device):
             weakref.proxy(self))
         self.attn_mask_builder = AttentionMaskBuilder(
             min(self.model_config.max_model_len,
-                int(os.getenv("PAGED_ATTENTION_MASK_LEN", 10000))), self.dtype)
+                int(os.getenv("PAGED_ATTENTION_MASK_LEN", 128))), self.dtype)
 
         # Set up speculative decoding.
         self.use_aux_hidden_state_outputs = False
@@ -834,9 +834,11 @@ def _make_attention_mask(self, seq_lens, query_lens, position,
                 seq_lens, query_lens, position, self.dtype, self.device)
         # Prefill without cache situation.
         elif attn_state == AscendAttentionState.PrefillNoCache:
-            max_seq_len = max(seq_lens, default=0)
+            # Note: `torch_npu._npu_flash_attention` only requires a 128x128 mask, so we hardcode it here.
+            # Once a new attention operator for prefill-only state is added,
+            # the mask generation logic here must be updated according to the new operator used.
             return self.attn_mask_builder.get_attn_mask(
-                max_seq_len, self.dtype, self.device)
+                128, self.dtype, self.device)
         # Prefill with cache hit.
         elif attn_state == AscendAttentionState.PrefillCacheHit:
             return self.attn_mask_builder.get_attn_mask(