mistralai 1.10.0__py3-none-any.whl → 1.11.1__py3-none-any.whl

mistralai_gcp/chat.py

@@ -5,7 +5,8 @@ from mistralai_gcp import models, utils
 from mistralai_gcp._hooks import HookContext
 from mistralai_gcp.types import OptionalNullable, UNSET
 from mistralai_gcp.utils import eventstreaming
-from typing import Any, List, Mapping, Optional, Union
+from mistralai_gcp.utils.unmarshal_json_response import unmarshal_json_response
+from typing import Any, Dict, List, Mapping, Optional, Union
 
 
 class Chat(BaseSDK):
@@ -22,6 +23,7 @@ class Chat(BaseSDK):
         stream: Optional[bool] = True,
         stop: Optional[Union[models.Stop, models.StopTypedDict]] = None,
         random_seed: OptionalNullable[int] = UNSET,
+        metadata: OptionalNullable[Dict[str, Any]] = UNSET,
         response_format: Optional[
             Union[models.ResponseFormat, models.ResponseFormatTypedDict]
         ] = None,
@@ -41,11 +43,12 @@ class Chat(BaseSDK):
             Union[models.Prediction, models.PredictionTypedDict]
         ] = None,
         parallel_tool_calls: Optional[bool] = None,
+        prompt_mode: OptionalNullable[models.MistralPromptMode] = UNSET,
         retries: OptionalNullable[utils.RetryConfig] = UNSET,
         server_url: Optional[str] = None,
         timeout_ms: Optional[int] = None,
         http_headers: Optional[Mapping[str, str]] = None,
-    ) -> Optional[eventstreaming.EventStream[models.CompletionEvent]]:
+    ) -> eventstreaming.EventStream[models.CompletionEvent]:
         r"""Stream chat completion
 
         Mistral AI provides the ability to stream responses back to a client in order to allow partial results for certain requests. Tokens will be sent as data-only server-sent events as they become available, with the stream terminated by a data: [DONE] message. Otherwise, the server will hold the request open until the timeout or until completion, with the response containing the full result as JSON.
@@ -58,14 +61,16 @@ class Chat(BaseSDK):
         :param stream:
         :param stop: Stop generation if this token is detected. Or if one of these tokens is detected when providing an array
         :param random_seed: The seed to use for random sampling. If set, different calls will generate deterministic results.
-        :param response_format:
-        :param tools:
-        :param tool_choice:
-        :param presence_penalty: presence_penalty determines how much the model penalizes the repetition of words or phrases. A higher presence penalty encourages the model to use a wider variety of words and phrases, making the output more diverse and creative.
-        :param frequency_penalty: frequency_penalty penalizes the repetition of words based on their frequency in the generated text. A higher frequency penalty discourages the model from repeating words that have already appeared frequently in the output, promoting diversity and reducing repetition.
+        :param metadata:
+        :param response_format: Specify the format that the model must output. By default it will use `{ \"type\": \"text\" }`. Setting to `{ \"type\": \"json_object\" }` enables JSON mode, which guarantees the message the model generates is in JSON. When using JSON mode you MUST also instruct the model to produce JSON yourself with a system or a user message. Setting to `{ \"type\": \"json_schema\" }` enables JSON schema mode, which guarantees the message the model generates is in JSON and follows the schema you provide.
+        :param tools: A list of tools the model may call. Use this to provide a list of functions the model may generate JSON inputs for.
+        :param tool_choice: Controls which (if any) tool is called by the model. `none` means the model will not call any tool and instead generates a message. `auto` means the model can pick between generating a message or calling one or more tools. `any` or `required` means the model must call one or more tools. Specifying a particular tool via `{\"type\": \"function\", \"function\": {\"name\": \"my_function\"}}` forces the model to call that tool.
+        :param presence_penalty: The `presence_penalty` determines how much the model penalizes the repetition of words or phrases. A higher presence penalty encourages the model to use a wider variety of words and phrases, making the output more diverse and creative.
+        :param frequency_penalty: The `frequency_penalty` penalizes the repetition of words based on their frequency in the generated text. A higher frequency penalty discourages the model from repeating words that have already appeared frequently in the output, promoting diversity and reducing repetition.
         :param n: Number of completions to return for each request, input tokens are only billed once.
-        :param prediction:
-        :param parallel_tool_calls:
+        :param prediction: Enable users to specify an expected completion, optimizing response times by leveraging known or predictable content.
+        :param parallel_tool_calls: Whether to enable parallel function calling during tool use, when enabled the model can call multiple tools in parallel.
+        :param prompt_mode: Allows toggling between the reasoning mode and no system prompt. When set to `reasoning` the system prompt for reasoning models will be used.
         :param retries: Override the default retry configuration for this method
         :param server_url: Override the default server URL for this method
         :param timeout_ms: Override the default request timeout configuration for this method in milliseconds
@@ -89,6 +94,7 @@ class Chat(BaseSDK):
             stream=stream,
             stop=stop,
             random_seed=random_seed,
+            metadata=metadata,
             messages=utils.get_pydantic_model(messages, List[models.Messages]),
             response_format=utils.get_pydantic_model(
                 response_format, Optional[models.ResponseFormat]
@@ -104,6 +110,7 @@ class Chat(BaseSDK):
                 prediction, Optional[models.Prediction]
             ),
             parallel_tool_calls=parallel_tool_calls,
+            prompt_mode=prompt_mode,
         )
 
         req = self._build_request(
@@ -122,6 +129,7 @@ class Chat(BaseSDK):
             get_serialized_body=lambda: utils.serialize_request_body(
                 request, False, False, "json", models.ChatCompletionStreamRequest
             ),
+            allow_empty_value=None,
             timeout_ms=timeout_ms,
         )
 
@@ -135,9 +143,10 @@ class Chat(BaseSDK):
 
         http_res = self.do_request(
             hook_ctx=HookContext(
+                config=self.sdk_configuration,
                 base_url=base_url or "",
                 operation_id="stream_chat",
-                oauth2_scopes=[],
+                oauth2_scopes=None,
                 security_source=self.sdk_configuration.security,
             ),
             request=req,
@@ -152,32 +161,23 @@ class Chat(BaseSDK):
                 http_res,
                 lambda raw: utils.unmarshal_json(raw, models.CompletionEvent),
                 sentinel="[DONE]",
+                client_ref=self,
             )
         if utils.match_response(http_res, "422", "application/json"):
             http_res_text = utils.stream_to_text(http_res)
-            response_data = utils.unmarshal_json(
-                http_res_text, models.HTTPValidationErrorData
+            response_data = unmarshal_json_response(
+                models.HTTPValidationErrorData, http_res, http_res_text
             )
-            raise models.HTTPValidationError(data=response_data)
+            raise models.HTTPValidationError(response_data, http_res, http_res_text)
         if utils.match_response(http_res, "4XX", "*"):
             http_res_text = utils.stream_to_text(http_res)
-            raise models.SDKError(
-                "API error occurred", http_res.status_code, http_res_text, http_res
-            )
+            raise models.SDKError("API error occurred", http_res, http_res_text)
         if utils.match_response(http_res, "5XX", "*"):
             http_res_text = utils.stream_to_text(http_res)
-            raise models.SDKError(
-                "API error occurred", http_res.status_code, http_res_text, http_res
-            )
+            raise models.SDKError("API error occurred", http_res, http_res_text)
 
-        content_type = http_res.headers.get("Content-Type")
         http_res_text = utils.stream_to_text(http_res)
-        raise models.SDKError(
-            f"Unexpected response received (code: {http_res.status_code}, type: {content_type})",
-            http_res.status_code,
-            http_res_text,
-            http_res,
-        )
+        raise models.SDKError("Unexpected response received", http_res, http_res_text)
 
     async def stream_async(
         self,
@@ -190,6 +190,7 @@ class Chat(BaseSDK):
         stream: Optional[bool] = True,
         stop: Optional[Union[models.Stop, models.StopTypedDict]] = None,
         random_seed: OptionalNullable[int] = UNSET,
+        metadata: OptionalNullable[Dict[str, Any]] = UNSET,
         response_format: Optional[
             Union[models.ResponseFormat, models.ResponseFormatTypedDict]
         ] = None,
@@ -209,11 +210,12 @@ class Chat(BaseSDK):
             Union[models.Prediction, models.PredictionTypedDict]
         ] = None,
         parallel_tool_calls: Optional[bool] = None,
+        prompt_mode: OptionalNullable[models.MistralPromptMode] = UNSET,
         retries: OptionalNullable[utils.RetryConfig] = UNSET,
         server_url: Optional[str] = None,
         timeout_ms: Optional[int] = None,
         http_headers: Optional[Mapping[str, str]] = None,
-    ) -> Optional[eventstreaming.EventStreamAsync[models.CompletionEvent]]:
+    ) -> eventstreaming.EventStreamAsync[models.CompletionEvent]:
         r"""Stream chat completion
 
         Mistral AI provides the ability to stream responses back to a client in order to allow partial results for certain requests. Tokens will be sent as data-only server-sent events as they become available, with the stream terminated by a data: [DONE] message. Otherwise, the server will hold the request open until the timeout or until completion, with the response containing the full result as JSON.
@@ -226,14 +228,16 @@ class Chat(BaseSDK):
         :param stream:
         :param stop: Stop generation if this token is detected. Or if one of these tokens is detected when providing an array
         :param random_seed: The seed to use for random sampling. If set, different calls will generate deterministic results.
-        :param response_format:
-        :param tools:
-        :param tool_choice:
-        :param presence_penalty: presence_penalty determines how much the model penalizes the repetition of words or phrases. A higher presence penalty encourages the model to use a wider variety of words and phrases, making the output more diverse and creative.
-        :param frequency_penalty: frequency_penalty penalizes the repetition of words based on their frequency in the generated text. A higher frequency penalty discourages the model from repeating words that have already appeared frequently in the output, promoting diversity and reducing repetition.
+        :param metadata:
+        :param response_format: Specify the format that the model must output. By default it will use `{ \"type\": \"text\" }`. Setting to `{ \"type\": \"json_object\" }` enables JSON mode, which guarantees the message the model generates is in JSON. When using JSON mode you MUST also instruct the model to produce JSON yourself with a system or a user message. Setting to `{ \"type\": \"json_schema\" }` enables JSON schema mode, which guarantees the message the model generates is in JSON and follows the schema you provide.
+        :param tools: A list of tools the model may call. Use this to provide a list of functions the model may generate JSON inputs for.
+        :param tool_choice: Controls which (if any) tool is called by the model. `none` means the model will not call any tool and instead generates a message. `auto` means the model can pick between generating a message or calling one or more tools. `any` or `required` means the model must call one or more tools. Specifying a particular tool via `{\"type\": \"function\", \"function\": {\"name\": \"my_function\"}}` forces the model to call that tool.
+        :param presence_penalty: The `presence_penalty` determines how much the model penalizes the repetition of words or phrases. A higher presence penalty encourages the model to use a wider variety of words and phrases, making the output more diverse and creative.
+        :param frequency_penalty: The `frequency_penalty` penalizes the repetition of words based on their frequency in the generated text. A higher frequency penalty discourages the model from repeating words that have already appeared frequently in the output, promoting diversity and reducing repetition.
         :param n: Number of completions to return for each request, input tokens are only billed once.
-        :param prediction:
-        :param parallel_tool_calls:
+        :param prediction: Enable users to specify an expected completion, optimizing response times by leveraging known or predictable content.
+        :param parallel_tool_calls: Whether to enable parallel function calling during tool use, when enabled the model can call multiple tools in parallel.
+        :param prompt_mode: Allows toggling between the reasoning mode and no system prompt. When set to `reasoning` the system prompt for reasoning models will be used.
         :param retries: Override the default retry configuration for this method
         :param server_url: Override the default server URL for this method
         :param timeout_ms: Override the default request timeout configuration for this method in milliseconds
@@ -257,6 +261,7 @@ class Chat(BaseSDK):
             stream=stream,
             stop=stop,
             random_seed=random_seed,
+            metadata=metadata,
             messages=utils.get_pydantic_model(messages, List[models.Messages]),
             response_format=utils.get_pydantic_model(
                 response_format, Optional[models.ResponseFormat]
@@ -272,6 +277,7 @@ class Chat(BaseSDK):
                 prediction, Optional[models.Prediction]
             ),
             parallel_tool_calls=parallel_tool_calls,
+            prompt_mode=prompt_mode,
         )
 
         req = self._build_request_async(
@@ -290,6 +296,7 @@ class Chat(BaseSDK):
             get_serialized_body=lambda: utils.serialize_request_body(
                 request, False, False, "json", models.ChatCompletionStreamRequest
             ),
+            allow_empty_value=None,
             timeout_ms=timeout_ms,
         )
 
@@ -303,9 +310,10 @@ class Chat(BaseSDK):
 
         http_res = await self.do_request_async(
             hook_ctx=HookContext(
+                config=self.sdk_configuration,
                 base_url=base_url or "",
                 operation_id="stream_chat",
-                oauth2_scopes=[],
+                oauth2_scopes=None,
                 security_source=self.sdk_configuration.security,
             ),
             request=req,
@@ -320,32 +328,23 @@ class Chat(BaseSDK):
                 http_res,
                 lambda raw: utils.unmarshal_json(raw, models.CompletionEvent),
                 sentinel="[DONE]",
+                client_ref=self,
             )
         if utils.match_response(http_res, "422", "application/json"):
             http_res_text = await utils.stream_to_text_async(http_res)
-            response_data = utils.unmarshal_json(
-                http_res_text, models.HTTPValidationErrorData
+            response_data = unmarshal_json_response(
+                models.HTTPValidationErrorData, http_res, http_res_text
             )
-            raise models.HTTPValidationError(data=response_data)
+            raise models.HTTPValidationError(response_data, http_res, http_res_text)
         if utils.match_response(http_res, "4XX", "*"):
             http_res_text = await utils.stream_to_text_async(http_res)
-            raise models.SDKError(
-                "API error occurred", http_res.status_code, http_res_text, http_res
-            )
+            raise models.SDKError("API error occurred", http_res, http_res_text)
         if utils.match_response(http_res, "5XX", "*"):
             http_res_text = await utils.stream_to_text_async(http_res)
-            raise models.SDKError(
-                "API error occurred", http_res.status_code, http_res_text, http_res
-            )
+            raise models.SDKError("API error occurred", http_res, http_res_text)
 
-        content_type = http_res.headers.get("Content-Type")
         http_res_text = await utils.stream_to_text_async(http_res)
-        raise models.SDKError(
-            f"Unexpected response received (code: {http_res.status_code}, type: {content_type})",
-            http_res.status_code,
-            http_res_text,
-            http_res,
-        )
+        raise models.SDKError("Unexpected response received", http_res, http_res_text)
 
     def complete(
         self,
@@ -366,6 +365,7 @@ class Chat(BaseSDK):
             ]
         ] = None,
         random_seed: OptionalNullable[int] = UNSET,
+        metadata: OptionalNullable[Dict[str, Any]] = UNSET,
         response_format: Optional[
             Union[models.ResponseFormat, models.ResponseFormatTypedDict]
         ] = None,
@@ -385,11 +385,12 @@ class Chat(BaseSDK):
             Union[models.Prediction, models.PredictionTypedDict]
         ] = None,
         parallel_tool_calls: Optional[bool] = None,
+        prompt_mode: OptionalNullable[models.MistralPromptMode] = UNSET,
         retries: OptionalNullable[utils.RetryConfig] = UNSET,
         server_url: Optional[str] = None,
         timeout_ms: Optional[int] = None,
         http_headers: Optional[Mapping[str, str]] = None,
-    ) -> Optional[models.ChatCompletionResponse]:
+    ) -> models.ChatCompletionResponse:
         r"""Chat Completion
 
         :param model: ID of the model to use. You can use the [List Available Models](/api/#tag/models/operation/list_models_v1_models_get) API to see all of your available models, or see our [Model overview](/models) for model descriptions.
@@ -400,14 +401,16 @@ class Chat(BaseSDK):
         :param stream: Whether to stream back partial progress. If set, tokens will be sent as data-only server-side events as they become available, with the stream terminated by a data: [DONE] message. Otherwise, the server will hold the request open until the timeout or until completion, with the response containing the full result as JSON.
         :param stop: Stop generation if this token is detected. Or if one of these tokens is detected when providing an array
         :param random_seed: The seed to use for random sampling. If set, different calls will generate deterministic results.
-        :param response_format:
-        :param tools:
-        :param tool_choice:
-        :param presence_penalty: presence_penalty determines how much the model penalizes the repetition of words or phrases. A higher presence penalty encourages the model to use a wider variety of words and phrases, making the output more diverse and creative.
-        :param frequency_penalty: frequency_penalty penalizes the repetition of words based on their frequency in the generated text. A higher frequency penalty discourages the model from repeating words that have already appeared frequently in the output, promoting diversity and reducing repetition.
+        :param metadata:
+        :param response_format: Specify the format that the model must output. By default it will use `{ \"type\": \"text\" }`. Setting to `{ \"type\": \"json_object\" }` enables JSON mode, which guarantees the message the model generates is in JSON. When using JSON mode you MUST also instruct the model to produce JSON yourself with a system or a user message. Setting to `{ \"type\": \"json_schema\" }` enables JSON schema mode, which guarantees the message the model generates is in JSON and follows the schema you provide.
+        :param tools: A list of tools the model may call. Use this to provide a list of functions the model may generate JSON inputs for.
+        :param tool_choice: Controls which (if any) tool is called by the model. `none` means the model will not call any tool and instead generates a message. `auto` means the model can pick between generating a message or calling one or more tools. `any` or `required` means the model must call one or more tools. Specifying a particular tool via `{\"type\": \"function\", \"function\": {\"name\": \"my_function\"}}` forces the model to call that tool.
+        :param presence_penalty: The `presence_penalty` determines how much the model penalizes the repetition of words or phrases. A higher presence penalty encourages the model to use a wider variety of words and phrases, making the output more diverse and creative.
+        :param frequency_penalty: The `frequency_penalty` penalizes the repetition of words based on their frequency in the generated text. A higher frequency penalty discourages the model from repeating words that have already appeared frequently in the output, promoting diversity and reducing repetition.
         :param n: Number of completions to return for each request, input tokens are only billed once.
-        :param prediction:
-        :param parallel_tool_calls:
+        :param prediction: Enable users to specify an expected completion, optimizing response times by leveraging known or predictable content.
+        :param parallel_tool_calls: Whether to enable parallel function calling during tool use, when enabled the model can call multiple tools in parallel.
+        :param prompt_mode: Allows toggling between the reasoning mode and no system prompt. When set to `reasoning` the system prompt for reasoning models will be used.
         :param retries: Override the default retry configuration for this method
         :param server_url: Override the default server URL for this method
         :param timeout_ms: Override the default request timeout configuration for this method in milliseconds
@@ -431,6 +434,7 @@ class Chat(BaseSDK):
             stream=stream,
             stop=stop,
             random_seed=random_seed,
+            metadata=metadata,
             messages=utils.get_pydantic_model(
                 messages, List[models.ChatCompletionRequestMessages]
             ),
@@ -448,6 +452,7 @@ class Chat(BaseSDK):
                 prediction, Optional[models.Prediction]
             ),
             parallel_tool_calls=parallel_tool_calls,
+            prompt_mode=prompt_mode,
         )
 
         req = self._build_request(
@@ -466,6 +471,7 @@ class Chat(BaseSDK):
             get_serialized_body=lambda: utils.serialize_request_body(
                 request, False, False, "json", models.ChatCompletionRequest
             ),
+            allow_empty_value=None,
             timeout_ms=timeout_ms,
         )
 
@@ -479,9 +485,10 @@ class Chat(BaseSDK):
 
         http_res = self.do_request(
             hook_ctx=HookContext(
+                config=self.sdk_configuration,
                 base_url=base_url or "",
                 operation_id="chat_completion_v1_chat_completions_post",
-                oauth2_scopes=[],
+                oauth2_scopes=None,
                 security_source=self.sdk_configuration.security,
             ),
             request=req,
@@ -491,33 +498,20 @@ class Chat(BaseSDK):
 
         response_data: Any = None
         if utils.match_response(http_res, "200", "application/json"):
-            return utils.unmarshal_json(
-                http_res.text, Optional[models.ChatCompletionResponse]
-            )
+            return unmarshal_json_response(models.ChatCompletionResponse, http_res)
         if utils.match_response(http_res, "422", "application/json"):
-            response_data = utils.unmarshal_json(
-                http_res.text, models.HTTPValidationErrorData
+            response_data = unmarshal_json_response(
+                models.HTTPValidationErrorData, http_res
             )
-            raise models.HTTPValidationError(data=response_data)
+            raise models.HTTPValidationError(response_data, http_res)
         if utils.match_response(http_res, "4XX", "*"):
             http_res_text = utils.stream_to_text(http_res)
-            raise models.SDKError(
-                "API error occurred", http_res.status_code, http_res_text, http_res
-            )
+            raise models.SDKError("API error occurred", http_res, http_res_text)
         if utils.match_response(http_res, "5XX", "*"):
             http_res_text = utils.stream_to_text(http_res)
-            raise models.SDKError(
-                "API error occurred", http_res.status_code, http_res_text, http_res
-            )
+            raise models.SDKError("API error occurred", http_res, http_res_text)
 
-        content_type = http_res.headers.get("Content-Type")
-        http_res_text = utils.stream_to_text(http_res)
-        raise models.SDKError(
-            f"Unexpected response received (code: {http_res.status_code}, type: {content_type})",
-            http_res.status_code,
-            http_res_text,
-            http_res,
-        )
+        raise models.SDKError("Unexpected response received", http_res)
 
     async def complete_async(
         self,
@@ -538,6 +532,7 @@ class Chat(BaseSDK):
             ]
         ] = None,
         random_seed: OptionalNullable[int] = UNSET,
+        metadata: OptionalNullable[Dict[str, Any]] = UNSET,
         response_format: Optional[
             Union[models.ResponseFormat, models.ResponseFormatTypedDict]
         ] = None,
@@ -557,11 +552,12 @@ class Chat(BaseSDK):
             Union[models.Prediction, models.PredictionTypedDict]
         ] = None,
         parallel_tool_calls: Optional[bool] = None,
+        prompt_mode: OptionalNullable[models.MistralPromptMode] = UNSET,
         retries: OptionalNullable[utils.RetryConfig] = UNSET,
         server_url: Optional[str] = None,
         timeout_ms: Optional[int] = None,
         http_headers: Optional[Mapping[str, str]] = None,
-    ) -> Optional[models.ChatCompletionResponse]:
+    ) -> models.ChatCompletionResponse:
         r"""Chat Completion
 
         :param model: ID of the model to use. You can use the [List Available Models](/api/#tag/models/operation/list_models_v1_models_get) API to see all of your available models, or see our [Model overview](/models) for model descriptions.
@@ -572,14 +568,16 @@ class Chat(BaseSDK):
         :param stream: Whether to stream back partial progress. If set, tokens will be sent as data-only server-side events as they become available, with the stream terminated by a data: [DONE] message. Otherwise, the server will hold the request open until the timeout or until completion, with the response containing the full result as JSON.
         :param stop: Stop generation if this token is detected. Or if one of these tokens is detected when providing an array
         :param random_seed: The seed to use for random sampling. If set, different calls will generate deterministic results.
-        :param response_format:
-        :param tools:
-        :param tool_choice:
-        :param presence_penalty: presence_penalty determines how much the model penalizes the repetition of words or phrases. A higher presence penalty encourages the model to use a wider variety of words and phrases, making the output more diverse and creative.
-        :param frequency_penalty: frequency_penalty penalizes the repetition of words based on their frequency in the generated text. A higher frequency penalty discourages the model from repeating words that have already appeared frequently in the output, promoting diversity and reducing repetition.
+        :param metadata:
+        :param response_format: Specify the format that the model must output. By default it will use `{ \"type\": \"text\" }`. Setting to `{ \"type\": \"json_object\" }` enables JSON mode, which guarantees the message the model generates is in JSON. When using JSON mode you MUST also instruct the model to produce JSON yourself with a system or a user message. Setting to `{ \"type\": \"json_schema\" }` enables JSON schema mode, which guarantees the message the model generates is in JSON and follows the schema you provide.
+        :param tools: A list of tools the model may call. Use this to provide a list of functions the model may generate JSON inputs for.
+        :param tool_choice: Controls which (if any) tool is called by the model. `none` means the model will not call any tool and instead generates a message. `auto` means the model can pick between generating a message or calling one or more tools. `any` or `required` means the model must call one or more tools. Specifying a particular tool via `{\"type\": \"function\", \"function\": {\"name\": \"my_function\"}}` forces the model to call that tool.
+        :param presence_penalty: The `presence_penalty` determines how much the model penalizes the repetition of words or phrases. A higher presence penalty encourages the model to use a wider variety of words and phrases, making the output more diverse and creative.
+        :param frequency_penalty: The `frequency_penalty` penalizes the repetition of words based on their frequency in the generated text. A higher frequency penalty discourages the model from repeating words that have already appeared frequently in the output, promoting diversity and reducing repetition.
         :param n: Number of completions to return for each request, input tokens are only billed once.
-        :param prediction:
-        :param parallel_tool_calls:
+        :param prediction: Enable users to specify an expected completion, optimizing response times by leveraging known or predictable content.
+        :param parallel_tool_calls: Whether to enable parallel function calling during tool use, when enabled the model can call multiple tools in parallel.
+        :param prompt_mode: Allows toggling between the reasoning mode and no system prompt. When set to `reasoning` the system prompt for reasoning models will be used.
         :param retries: Override the default retry configuration for this method
         :param server_url: Override the default server URL for this method
         :param timeout_ms: Override the default request timeout configuration for this method in milliseconds
@@ -603,6 +601,7 @@ class Chat(BaseSDK):
             stream=stream,
             stop=stop,
             random_seed=random_seed,
+            metadata=metadata,
             messages=utils.get_pydantic_model(
                 messages, List[models.ChatCompletionRequestMessages]
             ),
@@ -620,6 +619,7 @@ class Chat(BaseSDK):
                 prediction, Optional[models.Prediction]
             ),
             parallel_tool_calls=parallel_tool_calls,
+            prompt_mode=prompt_mode,
         )
 
         req = self._build_request_async(
@@ -638,6 +638,7 @@ class Chat(BaseSDK):
             get_serialized_body=lambda: utils.serialize_request_body(
                 request, False, False, "json", models.ChatCompletionRequest
             ),
+            allow_empty_value=None,
             timeout_ms=timeout_ms,
         )
 
@@ -651,9 +652,10 @@ class Chat(BaseSDK):
 
         http_res = await self.do_request_async(
             hook_ctx=HookContext(
+                config=self.sdk_configuration,
                 base_url=base_url or "",
                 operation_id="chat_completion_v1_chat_completions_post",
-                oauth2_scopes=[],
+                oauth2_scopes=None,
                 security_source=self.sdk_configuration.security,
             ),
             request=req,
@@ -663,30 +665,17 @@ class Chat(BaseSDK):
 
         response_data: Any = None
         if utils.match_response(http_res, "200", "application/json"):
-            return utils.unmarshal_json(
-                http_res.text, Optional[models.ChatCompletionResponse]
-            )
+            return unmarshal_json_response(models.ChatCompletionResponse, http_res)
         if utils.match_response(http_res, "422", "application/json"):
-            response_data = utils.unmarshal_json(
-                http_res.text, models.HTTPValidationErrorData
+            response_data = unmarshal_json_response(
+                models.HTTPValidationErrorData, http_res
             )
-            raise models.HTTPValidationError(data=response_data)
+            raise models.HTTPValidationError(response_data, http_res)
         if utils.match_response(http_res, "4XX", "*"):
             http_res_text = await utils.stream_to_text_async(http_res)
-            raise models.SDKError(
-                "API error occurred", http_res.status_code, http_res_text, http_res
-            )
+            raise models.SDKError("API error occurred", http_res, http_res_text)
         if utils.match_response(http_res, "5XX", "*"):
             http_res_text = await utils.stream_to_text_async(http_res)
-            raise models.SDKError(
-                "API error occurred", http_res.status_code, http_res_text, http_res
-            )
+            raise models.SDKError("API error occurred", http_res, http_res_text)
 
-        content_type = http_res.headers.get("Content-Type")
-        http_res_text = await utils.stream_to_text_async(http_res)
-        raise models.SDKError(
-            f"Unexpected response received (code: {http_res.status_code}, type: {content_type})",
-            http_res.status_code,
-            http_res_text,
-            http_res,
-        )
+        raise models.SDKError("Unexpected response received", http_res)