diff --git a/metagpt/configs/llm_config.py b/metagpt/configs/llm_config.py
index 4ed741af3..36f5d7ae7 100644
--- a/metagpt/configs/llm_config.py
+++ b/metagpt/configs/llm_config.py
@@ -38,13 +38,18 @@ class LLMConfig(YamlModel):
     Optional Fields in pydantic: https://docs.pydantic.dev/latest/migration/#required-optional-and-nullable-fields
     """
 
-    api_key: str
+    api_key: str = "sk-"
     api_type: LLMType = LLMType.OPENAI
     base_url: str = "https://api.openai.com/v1"
     api_version: Optional[str] = None
 
     model: Optional[str] = None  # also stands for DEPLOYMENT_NAME
 
+    # For Cloud Service Provider like Baidu/ Alibaba
+    access_key: Optional[str] = None
+    secret_key: Optional[str] = None
+    endpoint: Optional[str] = None  # for self-deployed model on the cloud
+
     # For Spark(Xunfei), maybe remove later
     app_id: Optional[str] = None
     api_secret: Optional[str] = None
diff --git a/metagpt/provider/base_llm.py b/metagpt/provider/base_llm.py
index b144471b5..2f57b15aa 100644
--- a/metagpt/provider/base_llm.py
+++ b/metagpt/provider/base_llm.py
@@ -11,11 +11,12 @@ from abc import ABC, abstractmethod
 from typing import Optional, Union
 
 from openai import AsyncOpenAI
+from pydantic import BaseModel
 
 from metagpt.configs.llm_config import LLMConfig
 from metagpt.logs import logger
 from metagpt.schema import Message
-from metagpt.utils.cost_manager import CostManager
+from metagpt.utils.cost_manager import CostManager, Costs
 
 
 class BaseLLM(ABC):
@@ -67,6 +68,28 @@ class BaseLLM(ABC):
     def _default_system_msg(self):
         return self._system_msg(self.system_prompt)
 
+    def _update_costs(self, usage: Union[dict, BaseModel], model: str = None, local_calc_usage: bool = True):
+        """update each request's token cost
+        Args:
+            model (str): model name or in some scenarios called endpoint
+            local_calc_usage (bool): some models don't calculate usage, it will overwrite LLMConfig.calc_usage
+        """
+        calc_usage = self.config.calc_usage and local_calc_usage
+        model = model if model else self.model
+        usage = usage.model_dump() if isinstance(usage, BaseModel) else usage
+        if calc_usage and self.cost_manager:
+            try:
+                prompt_tokens = int(usage.get("prompt_tokens", 0))
+                completion_tokens = int(usage.get("completion_tokens", 0))
+                self.cost_manager.update_cost(prompt_tokens, completion_tokens, model)
+            except Exception as e:
+                logger.error(f"{self.__class__.__name__} updats costs failed! exp: {e}")
+
+    def get_costs(self) -> Costs:
+        if not self.cost_manager:
+            return Costs(0, 0, 0, 0)
+        return self.cost_manager.get_costs()
+
     async def aask(
         self,
         msg: str,
diff --git a/metagpt/utils/cost_manager.py b/metagpt/utils/cost_manager.py
index c3aa7323f..efff07ae1 100644
--- a/metagpt/utils/cost_manager.py
+++ b/metagpt/utils/cost_manager.py
@@ -29,6 +29,7 @@ class CostManager(BaseModel):
     total_budget: float = 0
     max_budget: float = 10.0
     total_cost: float = 0
+    token_costs: dict[str, dict[str, float]] = TOKEN_COSTS  # different model's token cost
 
     def update_cost(self, prompt_tokens, completion_tokens, model):
         """
@@ -46,7 +47,8 @@ class CostManager(BaseModel):
             return
 
         cost = (
-            prompt_tokens * TOKEN_COSTS[model]["prompt"] + completion_tokens * TOKEN_COSTS[model]["completion"]
+            prompt_tokens * self.token_costs[model]["prompt"]
+            + completion_tokens * self.token_costs[model]["completion"]
         ) / 1000
         self.total_cost += cost
         logger.info(