diff --git a/setup.py b/setup.py
index 57acaddf5a..93227fd97b 100644
--- a/setup.py
+++ b/setup.py
@@ -104,7 +104,7 @@
 
 # Flash 2 group kept for backwards compatibility
 extra_deps['gpu-flash2'] = [
-    'flash-attn==2.6.3',
+    'flash-attn==2.7.2.post1',
 ]
 
 extra_deps['gpu'] = copy.deepcopy(extra_deps['gpu-flash2'])