diff --git a/convert-generic.py b/convert-generic.py
index 2336559c6..a1829a856 100755
--- a/convert-generic.py
+++ b/convert-generic.py
@@ -28,12 +28,16 @@ else:
     # output in the same directory as the model by default
     fname_out = dir_model / f'ggml-model-{ftype_str[ftype]}.gguf'
 
+is_big_endian = False
+if args.bigendian is not None:
+    is_big_endian = True
+
 print(f"Loading model: {dir_model.name}")
 
 hparams = model.Model.load_hparams(dir_model)
 
 model_class = model.Model.from_model_architecture(hparams["architectures"][0])
-model_instance = model_class(dir_model, ftype, fname_out)
+model_instance = model_class(dir_model, ftype, fname_out, is_big_endian)
 
 print("Set model parameters")
 model_instance.set_gguf_parameters()
diff --git a/model.py b/model.py
index c8ae1d51a..d5b72882f 100644
--- a/model.py
+++ b/model.py
@@ -30,16 +30,18 @@ class SentencePieceTokenTypes(IntEnum):
 
 
 class Model:
-    def __init__(self, dir_model: Path, ftype: int, fname_out: Path):
+    def __init__(self, dir_model: Path, ftype: int, fname_out: Path, is_big_endian: bool):
         self.dir_model = dir_model
         self.ftype = ftype
         self.fname_out = fname_out
+        self.is_big_endian = is_big_endian
+        self.endianess = gguf.GGUFEndian.BIG if is_big_endian else gguf.GGUFEndian.LITTLE
         self.is_safetensors = self._is_model_safetensors()
         self.num_parts = Model.count_model_parts(self.dir_model, ".safetensors" if self.is_safetensors else ".bin")
         self.part_names = self._get_part_names()
         self.hparams = Model.load_hparams(self.dir_model)
         self.model_arch = self._get_model_architecture()
-        self.gguf_writer = gguf.GGUFWriter(fname_out, gguf.MODEL_ARCH_NAMES[self.model_arch])
+        self.gguf_writer = gguf.GGUFWriter(fname_out, gguf.MODEL_ARCH_NAMES[self.model_arch], endianess=self.endianess)
 
     def set_vocab(self):
         self._set_vocab_gpt2()
diff --git a/util.py b/util.py
index 3388638de..d44ae183f 100644
--- a/util.py
+++ b/util.py
@@ -20,4 +20,6 @@ def parse_args() -> argparse.Namespace:
         "ftype", type=int, choices=[0, 1], default=1, nargs='?',
         help="output format - use 0 for float32, 1 for float16",
     )
+    parser.add_argument("--bigendian", action="store_true", help="model is executed on big endian machine")
+
     return parser.parse_args()