diff --git a/modules/ai/chat.ts b/modules/ai/chat.ts
index c3b070f..a369dc7 100644
--- a/modules/ai/chat.ts
+++ b/modules/ai/chat.ts
@@ -6,7 +6,7 @@ import { createFailResponse } from '../backend-node';
 import { handleRequestError } from '../session/api-helper';
 
 import { countToken, countTokenForString } from './encoding';
-import { ChatModel, DEFAULT_MAX_TOKEN, MAX_TOKENS, ChatOptions, ErrorResponse } from './constants';
+import { ChatModel, DEFAULT_MAX_TOKEN, MAX_TOKENS, ChatOptions, ErrorResponse, ModelPointer } from './constants';
 import { ChatCompletion, ChatCompletionInStream } from './types';
 import { getChatCompletionSupport } from './platform';
 import { checkRequest, RequestControlError } from './request-control';
@@ -68,7 +68,7 @@ export async function chat(options: ChatOptions) {
 
     const content = {
       stream: true,
-      model,
+      model: ModelPointer[model] ?? model,
       user: support.user,
       // max_tokens: maxToken - token,
       ...other,
diff --git a/modules/ai/constants.ts b/modules/ai/constants.ts
index 628e1ce..915bcf1 100644
--- a/modules/ai/constants.ts
+++ b/modules/ai/constants.ts
@@ -98,6 +98,14 @@ export enum ChatModel {
   GPT_4_32K = 'gpt-4-32k',
 }
 
+/**
+ * 模型指针
+ */
+export const ModelPointer: Record<string, string> = {
+  [ChatModel.GPT_4]: 'gpt-4-1106-preview',
+  [ChatModel.GPT_4_32K]: 'gpt-4-1106-preview',
+};
+
 export const ALL_SUPPORTED_CHAT_MODEL = [
   ChatModel.GPT3_5_TURBO,
   ChatModel.GPT3_5_TURBO_16K,