diff --git a/src/diffusers/models/transformers/transformer_hunyuan_video.py b/src/diffusers/models/transformers/transformer_hunyuan_video.py
index 846104718b9a..b5ff734eee25 100644
--- a/src/diffusers/models/transformers/transformer_hunyuan_video.py
+++ b/src/diffusers/models/transformers/transformer_hunyuan_video.py
@@ -721,7 +721,8 @@ def forward(
 
         for i in range(batch_size):
             attention_mask[i, : effective_sequence_length[i]] = True
-        attention_mask = attention_mask.unsqueeze(1)  # [B, 1, N], for broadcasting across attention heads
+        # [B, 1, 1, N], for broadcasting across attention heads
+        attention_mask = attention_mask.unsqueeze(1).unsqueeze(1)
 
         # 4. Transformer blocks
         if torch.is_grad_enabled() and self.gradient_checkpointing: