diff --git a/metagpt/actions/action.py b/metagpt/actions/action.py
index a1ab5c2bc..20c052aa9 100644
--- a/metagpt/actions/action.py
+++ b/metagpt/actions/action.py
@@ -15,6 +15,7 @@ from pydantic import BaseModel, ConfigDict, Field, model_validator
 from metagpt.actions.action_node import ActionNode
 from metagpt.configs.models_config import ModelsConfig
 from metagpt.context_mixin import ContextMixin
+from metagpt.provider.llm_provider_registry import create_llm_instance
 from metagpt.schema import (
     CodePlanAndChangeContext,
     CodeSummarizeContext,
@@ -45,7 +46,9 @@ class Action(SerializationMixin, ContextMixin, BaseModel):
     def _update_private_llm(cls, data: Any) -> Any:
         config = ModelsConfig.default().get(data.llm_name_or_type)
         if config:
-            data.llm.config = config
+            llm = create_llm_instance(config)
+            llm.cost_manager = data.llm.cost_manager
+            data.llm = llm
         return data
 
     @property
diff --git a/tests/data/config/config2.yaml b/tests/data/config/config2.yaml
index e0cefd8ff..8c9fc0703 100644
--- a/tests/data/config/config2.yaml
+++ b/tests/data/config/config2.yaml
@@ -3,7 +3,7 @@ llm:
   base_url: "YOUR_gpt-3.5-turbo_BASE_URL"
   api_key: "YOUR_gpt-3.5-turbo_API_KEY"
   model: "gpt-3.5-turbo"  # or gpt-3.5-turbo
-  proxy: "YOUR_gpt-3.5-turbo_PROXY"  # for LLM API requests
+  # proxy: "YOUR_gpt-3.5-turbo_PROXY"  # for LLM API requests
   # timeout: 600 # Optional. If set to 0, default value is 300.
   # Details: https://azure.microsoft.com/en-us/pricing/details/cognitive-services/openai-service/
   pricing_plan: "" # Optional. Use for Azure LLM when its model name is not the same as OpenAI's
@@ -13,7 +13,7 @@ models:
     api_type: "openai"  # or azure / ollama / groq etc.
     base_url: "YOUR_MODEL_1_BASE_URL"
     api_key: "YOUR_MODEL_1_API_KEY"
-    proxy: "YOUR_MODEL_1_PROXY"  # for LLM API requests
+    # proxy: "YOUR_MODEL_1_PROXY"  # for LLM API requests
     # timeout: 600 # Optional. If set to 0, default value is 300.
     # Details: https://azure.microsoft.com/en-us/pricing/details/cognitive-services/openai-service/
     pricing_plan: "" # Optional. Use for Azure LLM when its model name is not the same as OpenAI's