diff --git a/llama_stack/providers/inline/inference/meta_reference/__init__.py b/llama_stack/providers/inline/inference/meta_reference/__init__.py
index 18dc61d4a..2ae3e7a16 100644
--- a/llama_stack/providers/inline/inference/meta_reference/__init__.py
+++ b/llama_stack/providers/inline/inference/meta_reference/__init__.py
@@ -16,5 +16,6 @@ async def get_provider_impl(
     from .inference import MetaReferenceInferenceImpl
 
     impl = MetaReferenceInferenceImpl(config)
+    await impl.initialize()
 
     return impl
diff --git a/llama_stack/providers/inline/inference/meta_reference/inference.py b/llama_stack/providers/inline/inference/meta_reference/inference.py
index f2354aebb..d89bb21f7 100644
--- a/llama_stack/providers/inline/inference/meta_reference/inference.py
+++ b/llama_stack/providers/inline/inference/meta_reference/inference.py
@@ -73,6 +73,9 @@ class MetaReferenceInferenceImpl(
         self.model_id = None
         self.llama_model = None
 
+    async def initialize(self) -> None:
+        pass
+
     async def load_model(self, model_id, llama_model) -> None:
         log.info(f"Loading model `{model_id}`")
         if self.config.create_distributed_process_group: