diff --git a/src/agentscope/models/dashscope_model.py b/src/agentscope/models/dashscope_model.py
index 5d53f4d55..a3ac23613 100644
--- a/src/agentscope/models/dashscope_model.py
+++ b/src/agentscope/models/dashscope_model.py
@@ -356,11 +356,13 @@ def format(
 
             # prompt1
             [
+                {
+                    "role": "system",
+                    "content": "You're a helpful assistant"
+                },
                 {
                     "role": "user",
                     "content": (
-                        "You're a helpful assistant\\n"
-                        "\\n"
                         "## Conversation History\\n"
                         "Bob: Hi, how can I help you?\\n"
                         "user: What's the date today?"
diff --git a/src/agentscope/models/litellm_model.py b/src/agentscope/models/litellm_model.py
index 948481ae2..24399ad58 100644
--- a/src/agentscope/models/litellm_model.py
+++ b/src/agentscope/models/litellm_model.py
@@ -322,11 +322,13 @@ def format(
 
             # prompt1
             [
+                {
+                    "role": "system",
+                    "content": "You're a helpful assistant"
+                },
                 {
                     "role": "user",
                     "content": (
-                        "You're a helpful assistant\\n"
-                        "\\n"
                         "## Conversation History\\n"
                         "Bob: Hi, how can I help you?\\n"
                         "user: What's the date today?"
diff --git a/src/agentscope/models/model.py b/src/agentscope/models/model.py
index 429d34d7a..0586d4c94 100644
--- a/src/agentscope/models/model.py
+++ b/src/agentscope/models/model.py
@@ -243,7 +243,7 @@ def format_for_common_chat_models(
         *args: Union[Msg, Sequence[Msg]],
     ) -> List[dict]:
         """A common format strategy for chat models, which will format the
-        input messages into a user message.
+        input messages into a system message (if provided) and a user message.
 
         Note this strategy maybe not suitable for all scenarios,
         and developers are encouraged to implement their own prompt
@@ -270,11 +270,13 @@ def format_for_common_chat_models(
 
             # prompt1
             [
+                {
+                    "role": "system",
+                    "content": "You're a helpful assistant"
+                },
                 {
                     "role": "user",
                     "content": (
-                        "You're a helpful assistant\\n"
-                        "\\n"
                         "## Conversation History\\n"
                         "Bob: Hi, how can I help you?\\n"
                         "user: What's the date today?"
@@ -340,11 +342,6 @@ def format_for_common_chat_models(
                 )
 
         content_components = []
-        # Add system prompt at the beginning if provided
-        if sys_prompt is not None:
-            if not sys_prompt.endswith("\n"):
-                sys_prompt += "\n"
-            content_components.append(sys_prompt)
 
         # The conversation history is added to the user message if not empty
         if len(dialogue) > 0:
@@ -357,6 +354,10 @@ def format_for_common_chat_models(
             },
         ]
 
+        # Add system prompt at the beginning if provided
+        if sys_prompt is not None:
+            messages = [{"role": "system", "content": sys_prompt}] + messages
+
         return messages
 
     def _save_model_invocation(
diff --git a/src/agentscope/models/ollama_model.py b/src/agentscope/models/ollama_model.py
index 0dec98e34..bcb6bc18b 100644
--- a/src/agentscope/models/ollama_model.py
+++ b/src/agentscope/models/ollama_model.py
@@ -290,10 +290,13 @@ def format(
         .. code-block:: python
 
             [
+                {
+                    "role": "system",
+                    "content": "You're a helpful assistant"
+                },
                 {
                     "role": "user",
                     "content": (
-                        "You're a helpful assistant\\n\\n"
                         "## Conversation History\\n"
                         "Bob: Hi, how can I help you?\\n"
                         "user: What's the date today?"
@@ -329,7 +332,8 @@ def format(
                 )
 
         # record dialog history as a list of strings
-        system_content_template = []
+        system_prompt = None
+        history_content_template = []
         dialogue = []
         # TODO: here we default the url links to images
         images = []
@@ -337,9 +341,6 @@ def format(
             if i == 0 and unit.role == "system":
                 # system prompt
                 system_prompt = _convert_to_str(unit.content)
-                if not system_prompt.endswith("\n"):
-                    system_prompt += "\n"
-                system_content_template.append(system_prompt)
             else:
                 # Merge all messages into a conversation history prompt
                 dialogue.append(
@@ -352,21 +353,28 @@ def format(
         if len(dialogue) != 0:
             dialogue_history = "\n".join(dialogue)
 
-            system_content_template.extend(
+            history_content_template.extend(
                 ["## Conversation History", dialogue_history],
             )
 
-        system_content = "\n".join(system_content_template)
+        history_content = "\n".join(history_content_template)
 
-        system_message = {
+        # The conversation history message
+        history_message = {
             "role": "user",
-            "content": system_content,
+            "content": history_content,
         }
 
         if len(images) != 0:
-            system_message["images"] = images
+            history_message["images"] = images
+
+        if system_prompt is None:
+            return [history_message]
 
-        return [system_message]
+        return [
+            {"role": "system", "content": system_prompt},
+            history_message,
+        ]
 
 
 class OllamaEmbeddingWrapper(OllamaWrapperBase):
diff --git a/src/agentscope/models/openai_model.py b/src/agentscope/models/openai_model.py
index e25fc9061..7d7ccc081 100644
--- a/src/agentscope/models/openai_model.py
+++ b/src/agentscope/models/openai_model.py
@@ -9,7 +9,6 @@
     Dict,
     Optional,
     Generator,
-    get_args,
 )
 
 from loguru import logger
@@ -474,7 +473,9 @@ def format(
         *args: Union[Msg, Sequence[Msg]],
     ) -> List[dict]:
         """Format the input string and dictionary into the format that
-        OpenAI Chat API required.
+        OpenAI Chat API required. If you're using a OpenAI-compatible model
+        without a prefix "gpt-" in its name, the format method will
+        automatically format the input messages into the required format.
 
         Args:
             args (`Union[Msg, Sequence[Msg]]`):
@@ -487,17 +488,9 @@ def format(
                 The formatted messages in the format that OpenAI Chat API
                 required.
         """
-        # Check if the OpenAI library is installed
-        try:
-            import openai
-        except ImportError as e:
-            raise ImportError(
-                "Cannot find openai package, please install it by "
-                "`pip install openai`",
-            ) from e
 
         # Format messages according to the model name
-        if self.model_name in get_args(openai.types.ChatModel):
+        if self.model_name.startswith("gpt-"):
             return OpenAIChatWrapper.static_format(
                 *args,
                 model_name=self.model_name,
diff --git a/src/agentscope/models/post_model.py b/src/agentscope/models/post_model.py
index fbd09bd0e..def456c2f 100644
--- a/src/agentscope/models/post_model.py
+++ b/src/agentscope/models/post_model.py
@@ -192,8 +192,9 @@ def format(
         self,
         *args: Union[Msg, Sequence[Msg]],
     ) -> Union[List[dict]]:
-        """Format the input messages into a list of dict, which is
-        compatible to OpenAI Chat API.
+        """Format the input messages into a list of dict according to the model
+        name. For example, if the model name is prefixed with "gpt-", the
+        input messages will be formatted for OpenAI models.
 
         Args:
             args (`Union[Msg, Sequence[Msg]]`):
diff --git a/src/agentscope/models/yi_model.py b/src/agentscope/models/yi_model.py
index 9d02dd17c..0acc314d2 100644
--- a/src/agentscope/models/yi_model.py
+++ b/src/agentscope/models/yi_model.py
@@ -230,11 +230,13 @@ def format(
 
             # prompt1
             [
+                {
+                    "role": "system",
+                    "content": "You're a helpful assistant"
+                },
                 {
                     "role": "user",
                     "content": (
-                        "You're a helpful assistant\\n"
-                        "\\n"
                         "## Conversation History\\n"
                         "Bob: Hi, how can I help you?\\n"
                         "user: What's the date today?"
diff --git a/src/agentscope/models/zhipu_model.py b/src/agentscope/models/zhipu_model.py
index c767dd5b4..27455ba9e 100644
--- a/src/agentscope/models/zhipu_model.py
+++ b/src/agentscope/models/zhipu_model.py
@@ -325,11 +325,13 @@ def format(
 
             # prompt1
             [
+                {
+                    "role": "system",
+                    "content": "You're a helpful assistant"
+                },
                 {
                     "role": "user",
                     "content": (
-                        "You're a helpful assistant\\n"
-                        "\\n"
                         "## Conversation History\\n"
                         "Bob: Hi, how can I help you?\\n"
                         "user: What's the date today?"
diff --git a/tests/format_test.py b/tests/format_test.py
index de0ccb510..217936ce0 100644
--- a/tests/format_test.py
+++ b/tests/format_test.py
@@ -238,11 +238,13 @@ def test_openai_chat_with_other_models(
 
         # correct format
         ground_truth = [
+            {
+                "role": "system",
+                "content": "You are a helpful assistant",
+            },
             {
                 "role": "user",
                 "content": (
-                    "You are a helpful assistant\n"
-                    "\n"
                     "## Conversation History\n"
                     "user: What is the weather today?\n"
                     "assistant: It is sunny today"
@@ -260,11 +262,13 @@ def test_format_for_common_models(self) -> None:
 
         # correct format
         ground_truth = [
+            {
+                "role": "system",
+                "content": "You are a helpful assistant",
+            },
             {
                 "role": "user",
                 "content": (
-                    "You are a helpful assistant\n"
-                    "\n"
                     "## Conversation History\n"
                     "user: What is the weather today?\n"
                     "assistant: It is sunny today"
@@ -282,11 +286,13 @@ def test_ollama_chat(self) -> None:
 
         # correct format
         ground_truth = [
+            {
+                "role": "system",
+                "content": "You are a helpful assistant",
+            },
             {
                 "role": "user",
                 "content": (
-                    "You are a helpful assistant\n"
-                    "\n"
                     "## Conversation History\n"
                     "user: What is the weather today?\n"
                     "assistant: It is sunny today"
@@ -358,10 +364,12 @@ def test_dashscope_chat(self) -> None:
         )
 
         ground_truth = [
+            {
+                "role": "system",
+                "content": "You are a helpful assistant",
+            },
             {
                 "content": (
-                    "You are a helpful assistant\n"
-                    "\n"
                     "## Conversation History\n"
                     "user: What is the weather today?\n"
                     "assistant: It is sunny today"
@@ -386,10 +394,12 @@ def test_zhipuai_chat(self) -> None:
         )
 
         ground_truth = [
+            {
+                "role": "system",
+                "content": "You are a helpful assistant",
+            },
             {
                 "content": (
-                    "You are a helpful assistant\n"
-                    "\n"
                     "## Conversation History\n"
                     "user: What is the weather today?\n"
                     "assistant: It is sunny today"
@@ -414,11 +424,13 @@ def test_litellm_chat(self) -> None:
         )
 
         ground_truth = [
+            {
+                "role": "system",
+                "content": "You are a helpful assistant",
+            },
             {
                 "role": "user",
                 "content": (
-                    "You are a helpful assistant\n"
-                    "\n"
                     "## Conversation History\n"
                     "user: What is the weather today?\n"
                     "assistant: It is sunny today"