diff --git a/gptqmodel/models/base.py b/gptqmodel/models/base.py
index 037b9c2a0..6f0d76d89 100644
--- a/gptqmodel/models/base.py
+++ b/gptqmodel/models/base.py
@@ -614,8 +614,7 @@ def save_quantized(
     ):
         """save quantized model and configs to local disk"""
         if self.from_quantized:
-            raise NotImplementedError("Saving a quantized model again is not supported. \n"
-                                      "If you need to shard the model file, refer to shard_quantized().")
+            raise NotImplementedError("Saving a loaded quantized model is not supported. If you need to re-shard the model, please use `GPTQModel.shard_quantized()` api.")
 
         os.makedirs(save_dir, exist_ok=True)