diff --git a/llama_stack/providers/inline/agents/meta_reference/agent_instance.py b/llama_stack/providers/inline/agents/meta_reference/agent_instance.py
index 3619b3f67..95c72d882 100644
--- a/llama_stack/providers/inline/agents/meta_reference/agent_instance.py
+++ b/llama_stack/providers/inline/agents/meta_reference/agent_instance.py
@@ -545,6 +545,7 @@ class ChatAgent(ShieldRunnerMixin):
                             )
 
                     elif delta.type == "text":
+                        delta.text = "hello"
                         content += delta.text
                         if stream and event.stop_reason is None:
                             yield AgentTurnResponseStreamChunk(
diff --git a/llama_stack/providers/remote/inference/passthrough/passthrough.py b/llama_stack/providers/remote/inference/passthrough/passthrough.py
index aa8a87bf7..6fe3deb8d 100644
--- a/llama_stack/providers/remote/inference/passthrough/passthrough.py
+++ b/llama_stack/providers/remote/inference/passthrough/passthrough.py
@@ -6,7 +6,7 @@
 
 from typing import AsyncGenerator, List, Optional
 
-from llama_stack_client import LlamaStackClient
+from llama_stack_client import AsyncLlamaStackClient
 
 from llama_stack.apis.common.content_types import InterleavedContent
 from llama_stack.apis.inference import (
@@ -46,7 +46,7 @@ class PassthroughInferenceAdapter(Inference):
     async def register_model(self, model: Model) -> Model:
         return model
 
-    def _get_client(self) -> LlamaStackClient:
+    def _get_client(self) -> AsyncLlamaStackClient:
         passthrough_url = None
         passthrough_api_key = None
         provider_data = None
@@ -71,7 +71,7 @@ class PassthroughInferenceAdapter(Inference):
                 )
             passthrough_api_key = provider_data.passthrough_api_key
 
-        return LlamaStackClient(
+        return AsyncLlamaStackClient(
             base_url=passthrough_url,
             api_key=passthrough_api_key,
             provider_data=provider_data,
@@ -103,7 +103,7 @@ class PassthroughInferenceAdapter(Inference):
         params = {key: value for key, value in params.items() if value is not None}
 
         # only pass through the not None params
-        return client.inference.completion(**params)
+        return await client.inference.completion(**params)
 
     async def chat_completion(
         self,
@@ -123,7 +123,7 @@ class PassthroughInferenceAdapter(Inference):
         client = self._get_client()
         model = await self.model_store.get_model(model_id)
 
-        params = {
+        reqeust_params = {
             "model_id": model.provider_resource_id,
             "messages": messages,
             "sampling_params": sampling_params,
@@ -134,11 +134,34 @@ class PassthroughInferenceAdapter(Inference):
             "stream": stream,
             "logprobs": logprobs,
         }
+        
 
-        params = {key: value for key, value in params.items() if value is not None}
+        request_params = {key: value for key, value in reqeust_params.items() if value is not None}
+
+        json_params = {}
+        from llama_stack.distribution.library_client import (
+            convert_pydantic_to_json_value,
+        )
+
+        # cast everything to json dict
+        for key, value in request_params.items():
+            json_input = convert_pydantic_to_json_value(value)
+            if isinstance(json_input, dict):
+                json_input = {k: v for k, v in json_input.items() if v is not None}
+            elif isinstance(json_input, list):
+                json_input = [x for x in json_input if x is not None]
+                new_input = []
+                for x in json_input:
+                    if isinstance(x, dict):
+                        x = {k: v for k, v in x.items() if v is not None}
+                    new_input.append(x)
+                json_input = new_input
+
+            # if key != "tools":
+            json_params[key] = json_input
 
         # only pass through the not None params
-        return client.inference.chat_completion(**params)
+        return await client.inference.chat_completion(**json_params)    
 
     async def embeddings(
         self,
@@ -151,7 +174,7 @@ class PassthroughInferenceAdapter(Inference):
         client = self._get_client()
         model = await self.model_store.get_model(model_id)
 
-        return client.inference.embeddings(
+        return await client.inference.embeddings(
             model_id=model.provider_resource_id,
             contents=contents,
             text_truncation=text_truncation,
diff --git a/llama_stack/templates/passthrough/run.yaml b/llama_stack/templates/passthrough/run.yaml
index 2548faa5d..195c8f115 100644
--- a/llama_stack/templates/passthrough/run.yaml
+++ b/llama_stack/templates/passthrough/run.yaml
@@ -20,6 +20,13 @@ providers:
   - provider_id: sentence-transformers
     provider_type: inline::sentence-transformers
     config: {}
+  - provider_id: meta-reference-inference
+    provider_type: inline::meta-reference
+    config:
+      model: meta-llama/Llama-Guard-3-1B
+      max_seq_len: 4096
+      checkpoint_dir: ${env.INFERENCE_CHECKPOINT_DIR:null}
+      # api_key: ${env.TOGETHER_API_KEY}
   vector_io:
   - provider_id: faiss
     provider_type: inline::faiss
@@ -103,8 +110,12 @@ models:
   provider_id: passthrough
   provider_model_id: llama3.2-11b-vision-instruct
   model_type: llm
+- metadata: {}
+  model_id: meta-llama/Llama-Guard-3-1B
+  provider_id: meta-reference-inference
+  model_type: llm
 shields:
-- shield_id: meta-llama/Llama-Guard-3-8B
+- shield_id: meta-llama/Llama-Guard-3-1B
 vector_dbs: []
 datasets: []
 scoring_fns: []