diff --git a/python/llm/src/bigdl/llm/ggml/model/bloom/bloom.py b/python/llm/src/bigdl/llm/ggml/model/bloom/bloom.py
index 0eea6528..ec33ddaa 100644
--- a/python/llm/src/bigdl/llm/ggml/model/bloom/bloom.py
+++ b/python/llm/src/bigdl/llm/ggml/model/bloom/bloom.py
@@ -313,9 +313,6 @@ class Bloom(GenerationMixin):
                     }
                 }
 
-    def free(self):
-        bloom_free(self.ctx)
-
     def _tokenize(self, text: bytes, add_bos: bool = False) -> List[int]:
         """Tokenize a string.
 
@@ -427,3 +424,8 @@ class Bloom(GenerationMixin):
                            seed=self.seed,
                            n_threads=self.n_threads,
                            n_batch=self.n_batch)
+
+    def __del__(self):
+        if self.ctx is not None:
+            bloom_free(self.ctx)
+            self.ctx = None
diff --git a/python/llm/src/bigdl/llm/ggml/model/starcoder/starcoder.py b/python/llm/src/bigdl/llm/ggml/model/starcoder/starcoder.py
index e40fc991..e54392d7 100644
--- a/python/llm/src/bigdl/llm/ggml/model/starcoder/starcoder.py
+++ b/python/llm/src/bigdl/llm/ggml/model/starcoder/starcoder.py
@@ -315,9 +315,6 @@ class Starcoder(GenerationMixin):
                         }
                     }
 
-    def free(self):
-        starcoder_free(self.ctx)
-
     def _tokenize(self, text: bytes, add_bos: bool = False) -> List[int]:
         """Tokenize a string.
 
@@ -433,3 +430,8 @@ class Starcoder(GenerationMixin):
                                seed=self.seed,
                                n_threads=self.n_threads,
                                n_batch=self.n_batch)
+
+    def __del__(self):
+        if self.ctx is not None:
+            starcoder_free(self.ctx)
+            self.ctx = None