diff --git a/jaxlib/triton/compat.py b/jaxlib/triton/compat.py
index 3c2125042636..0e5d948263ad 100644
--- a/jaxlib/triton/compat.py
+++ b/jaxlib/triton/compat.py
@@ -672,13 +672,13 @@ def max(x: tensor, y: tensor) -> tensor:
     assert x.shape == y.shape
     if x.dtype.is_floating():
       # TODO(slebedev): Triton promotes bfloat16 to float32 and back here.
-      return tensor(arith_dialect.maxnumf(x.handle, y.handle), x.dtype)
+      return tensor(arith_dialect.maxnumf(x.handle, y.handle), x.type)
     if not x.dtype.is_int():
       raise NotImplementedError(f"unsupported dtypes: {x.dtype} and {y.dtype}")
     elif x.dtype.is_int_signed():
-      return tensor(arith_dialect.maxsi(x.handle, y.handle), x.dtype)
+      return tensor(arith_dialect.maxsi(x.handle, y.handle), x.type)
     else:
-      return tensor(arith_dialect.maxui(x.handle, y.handle), x.dtype)
+      return tensor(arith_dialect.maxui(x.handle, y.handle), x.type)
 
   @staticmethod
   def min(x: tensor, y: tensor) -> tensor:
@@ -686,13 +686,13 @@ def min(x: tensor, y: tensor) -> tensor:
     assert x.shape == y.shape
     if x.dtype.is_floating():
       # TODO(slebedev): Triton promotes bfloat16 to float32 and back here.
-      return tensor(arith_dialect.minnumf(x.handle, y.handle), x.dtype)
+      return tensor(arith_dialect.minnumf(x.handle, y.handle), x.type)
     if not x.dtype.is_int():
       raise NotImplementedError(f"unsupported dtypes: {x.dtype} and {y.dtype}")
     elif x.dtype.is_int_signed():
-      return tensor(arith_dialect.minsi(x.handle, y.handle), x.dtype)
+      return tensor(arith_dialect.minsi(x.handle, y.handle), x.type)
     else:
-      return tensor(arith_dialect.minui(x.handle, y.handle), x.dtype)
+      return tensor(arith_dialect.minui(x.handle, y.handle), x.type)
 
   sin = libdevice_extern_elementwise({
       (float32,): ("__nv_sinf", float32),