diff --git a/apps/setting/models_provider/impl/aliyun_bai_lian_model_provider/model/tts.py b/apps/setting/models_provider/impl/aliyun_bai_lian_model_provider/model/tts.py
index cdcdd426c..35a46d189 100644
--- a/apps/setting/models_provider/impl/aliyun_bai_lian_model_provider/model/tts.py
+++ b/apps/setting/models_provider/impl/aliyun_bai_lian_model_provider/model/tts.py
@@ -17,12 +17,12 @@ class AliyunBaiLianTextToSpeech(MaxKBBaseModel, BaseTextToSpeech):
         super().__init__(**kwargs)
         self.api_key = kwargs.get('api_key')
         self.model = kwargs.get('model')
-        self.voice = kwargs.get('voice', 'longxiaochun')
-        self.speech_rate = kwargs.get('speech_rate', 1.0)
+        self.voice = kwargs.get('voice')
+        self.speech_rate = kwargs.get('speech_rate')
 
     @staticmethod
     def new_instance(model_type, model_name, model_credential: Dict[str, object], **model_kwargs):
-        optional_params = {}
+        optional_params = {'voice': 'longxiaochun', 'speech_rate': 1.0}
         if 'voice' in model_kwargs and model_kwargs['voice'] is not None:
             optional_params['voice'] = model_kwargs['voice']
         if 'speech_rate' in model_kwargs and model_kwargs['speech_rate'] is not None:
diff --git a/apps/setting/models_provider/impl/openai_model_provider/model/tts.py b/apps/setting/models_provider/impl/openai_model_provider/model/tts.py
index 2fa63d51d..839a59dc1 100644
--- a/apps/setting/models_provider/impl/openai_model_provider/model/tts.py
+++ b/apps/setting/models_provider/impl/openai_model_provider/model/tts.py
@@ -27,7 +27,7 @@ class OpenAITextToSpeech(MaxKBBaseModel, BaseTextToSpeech):
 
     @staticmethod
     def new_instance(model_type, model_name, model_credential: Dict[str, object], **model_kwargs):
-        optional_params = {}
+        optional_params = {'voice': 'alloy'}
         if 'voice' in model_kwargs and model_kwargs['voice'] is not None:
             optional_params['voice'] = model_kwargs['voice']
         return OpenAITextToSpeech(
diff --git a/apps/setting/models_provider/impl/volcanic_engine_model_provider/model/tts.py b/apps/setting/models_provider/impl/volcanic_engine_model_provider/model/tts.py
index 7a5af8e4b..5b438dc78 100644
--- a/apps/setting/models_provider/impl/volcanic_engine_model_provider/model/tts.py
+++ b/apps/setting/models_provider/impl/volcanic_engine_model_provider/model/tts.py
@@ -54,12 +54,12 @@ class VolcanicEngineTextToSpeech(MaxKBBaseModel, BaseTextToSpeech):
         self.volcanic_token = kwargs.get('volcanic_token')
         self.volcanic_app_id = kwargs.get('volcanic_app_id')
         self.volcanic_cluster = kwargs.get('volcanic_cluster')
-        self.voice_type = kwargs.get('voice_type', 'BV002_streaming')
-        self.speed_ratio = kwargs.get('speed_ratio', 1.0)
+        self.voice_type = kwargs.get('voice_type')
+        self.speed_ratio = kwargs.get('speed_ratio')
 
     @staticmethod
     def new_instance(model_type, model_name, model_credential: Dict[str, object], **model_kwargs):
-        optional_params = {}
+        optional_params = {'voice_type': 'BV002_streaming', 'speed_ratio': 1.0}
         if 'voice_type' in model_kwargs and model_kwargs['voice_type'] is not None:
             optional_params['voice_type'] = model_kwargs['voice_type']
         if 'speed_ratio' in model_kwargs and model_kwargs['speed_ratio'] is not None:
diff --git a/apps/setting/models_provider/impl/xf_model_provider/model/tts.py b/apps/setting/models_provider/impl/xf_model_provider/model/tts.py
index d6635c5dc..9bea1207a 100644
--- a/apps/setting/models_provider/impl/xf_model_provider/model/tts.py
+++ b/apps/setting/models_provider/impl/xf_model_provider/model/tts.py
@@ -46,12 +46,12 @@ class XFSparkTextToSpeech(MaxKBBaseModel, BaseTextToSpeech):
         self.spark_app_id = kwargs.get('spark_app_id')
         self.spark_api_key = kwargs.get('spark_api_key')
         self.spark_api_secret = kwargs.get('spark_api_secret')
-        self.vcn = kwargs.get('vcn', 'xiaoyan')
-        self.speed = kwargs.get('speed', 50)
+        self.vcn = kwargs.get('vcn')
+        self.speed = kwargs.get('speed')
 
     @staticmethod
     def new_instance(model_type, model_name, model_credential: Dict[str, object], **model_kwargs):
-        optional_params = {}
+        optional_params = {'vcn': 'xiaoyan', 'speed': 50}
         if 'vcn' in model_kwargs and model_kwargs['vcn'] is not None:
             optional_params['vcn'] = model_kwargs['vcn']
         if 'speed' in model_kwargs and model_kwargs['speed'] is not None:
diff --git a/apps/setting/models_provider/impl/xinference_model_provider/model/tts.py b/apps/setting/models_provider/impl/xinference_model_provider/model/tts.py
index bb8a3faed..bf3f91910 100644
--- a/apps/setting/models_provider/impl/xinference_model_provider/model/tts.py
+++ b/apps/setting/models_provider/impl/xinference_model_provider/model/tts.py
@@ -27,7 +27,7 @@ class XInferenceTextToSpeech(MaxKBBaseModel, BaseTextToSpeech):
 
     @staticmethod
     def new_instance(model_type, model_name, model_credential: Dict[str, object], **model_kwargs):
-        optional_params = {}
+        optional_params = {'voice': '中文女'}
         if 'voice' in model_kwargs and model_kwargs['voice'] is not None:
             optional_params['voice'] = model_kwargs['voice']
         return XInferenceTextToSpeech(