diff --git a/.env b/.env
index fb7ec9a7cd..f3e8c6856e 100644
--- a/.env
+++ b/.env
@@ -1,7 +1,7 @@
 # Glob API URL
 VITE_GLOB_API_URL=/api
 
-#VITE_APP_API_BASE_URL=http://127.0.0.1:3002 
+# VITE_APP_API_BASE_URL=http://127.0.0.1:3002 
 VITE_APP_API_BASE_URL=http://192.168.31.186:3002 
 
 # Whether long replies are supported, which may result in higher API fees
diff --git a/src/locales/en-US.ts b/src/locales/en-US.ts
index 2c9b4685f1..ad10753a32 100644
--- a/src/locales/en-US.ts
+++ b/src/locales/en-US.ts
@@ -303,6 +303,8 @@ export default {
     "historyToken": "More context improves accuracy but consumes more credits",
     "historyTCnt": "Reply Count",
     "historyTCntInfo": "Higher reply count may consume more credits",
+    "modelMaxTokens": "Max supported tokens",
+    "modelMaxTokensInfo": "The maximum amount supported by the model",
     "role": "Role Setting",
     "rolePlaceholder": "Set an exclusive role for your conversation (optional)",
     "loading2": "Loading...",
diff --git a/src/locales/fr-FR.ts b/src/locales/fr-FR.ts
index ff6f4019fa..6c549e3de4 100644
--- a/src/locales/fr-FR.ts
+++ b/src/locales/fr-FR.ts
@@ -301,6 +301,8 @@ export default {
         "historyToken": "Plus de contexte améliore la précision mais consomme plus de crédits",
         "historyTCnt": "Nombre de Réponses",
         "historyTCntInfo": "Un nombre plus élevé de réponses peut consommer plus de crédits",
+        "modelMaxTokens": "Le montant maximum pris en charge par le modèle",
+        "modelMaxTokensInfo": "Le montant maximum pris en charge par le modèle",
         "role": "Paramètre de Rôle",
         "rolePlaceholder": "Définissez un rôle exclusif pour votre conversation (facultatif)",
         "loading2": "Chargement...",
diff --git a/src/locales/ko-KR.ts b/src/locales/ko-KR.ts
index 504a153c5a..0914a50909 100644
--- a/src/locales/ko-KR.ts
+++ b/src/locales/ko-KR.ts
@@ -300,6 +300,8 @@ export default {
     ,"historyToken": "더 많은 컨텍스트는 기억을 더 정확하게 만들지만 더 많은 크레딧을 소비할 수 있습니다."
     ,"historyTCnt": "답장 수"
     ,"historyTCntInfo": "답장 수가 많을수록 더 많은 크레딧이 소비될 수 있습니다."
+    ,"modelMaxTokens": "모델이 지원하는 최대 한도"
+    ,"modelMaxTokensInfo": "모델이 지원하는 최대 한도"
     ,"role": "역할 설정"
     ,"rolePlaceholder": "대화에 고유한 역할을 설정하십시오. 필수는 아닙니다."
     ,"loading2": "로딩 중..."
diff --git a/src/locales/ru-RU.ts b/src/locales/ru-RU.ts
index 2bddb4241f..22ee9f58af 100644
--- a/src/locales/ru-RU.ts
+++ b/src/locales/ru-RU.ts
@@ -304,6 +304,8 @@ export default {
     "historyToken": "Больше контекста делает память точнее, но расходует больше квоты",
     "historyTCnt": "Количество ответов",
     "historyTCntInfo": "Больше ответов, возможно, потребуется больше квоты",
+    "modelMaxTokens": "Максимальный лимит, поддерживаемый моделью",
+    "modelMaxTokensInfo": "Максимальный лимит, поддерживаемый моделью",
     "role": "Настройка роли",
     "rolePlaceholder": "Дайте своему разговору уникальную роль, необязательно",
     "loading2": "Загрузка...",
diff --git a/src/locales/tr-TR.ts b/src/locales/tr-TR.ts
index a08bcc139f..686f89efa2 100644
--- a/src/locales/tr-TR.ts
+++ b/src/locales/tr-TR.ts
@@ -300,6 +300,8 @@ export default {
         "historyToken": "Daha fazla bağlam, doğruluğu artırır ancak daha fazla kredi tüketir",
         "historyTCnt": "Yanıt Sayısı",
         "historyTCntInfo": "Daha yüksek yanıt sayısı daha fazla kredi tüketebilir",
+        "modelMaxTokens": "Modelin desteklediği maksimum limit",
+        "modelMaxTokensInfo": "Modelin desteklediği maksimum limit",
         "role": "Rol Ayarı",
         "rolePlaceholder": "Konuşmanız için özel bir rol ayarlayın (isteğe bağlı)",
         "loading2": "Yükleniyor...",
diff --git a/src/locales/vi-VN.ts b/src/locales/vi-VN.ts
index de26c27d08..22877e278d 100644
--- a/src/locales/vi-VN.ts
+++ b/src/locales/vi-VN.ts
@@ -300,6 +300,8 @@ export default {
     "historyToken": "Số ngữ cảnh nhiều hơn sẽ làm cho bộ nhớ chính xác hơn, nhưng sẽ tiêu tốn nhiều chi phí hơn",
     "historyTCnt": "Số câu trả lời",
     "historyTCntInfo": "Số câu trả lời càng nhiều, khả năng tiêu tốn chi phí càng cao",
+    "modelMaxTokens": "Giới hạn tối đa mà mô hình hỗ trợ",
+    "modelMaxTokensInfo": "Giới hạn tối đa mà mô hình hỗ trợ",
     "role": "Đặt vai trò",
     "rolePlaceholder": "Đặt một vai trò riêng cho cuộc trò chuyện của bạn, không bắt buộc",
     "loading2": "Đang tải...",
diff --git a/src/locales/zh-CN.ts b/src/locales/zh-CN.ts
index ce07308595..3eab90db6c 100644
--- a/src/locales/zh-CN.ts
+++ b/src/locales/zh-CN.ts
@@ -134,6 +134,8 @@ export default {
     ,historyToken:'更多的上下文会使记忆更精确，但会消耗更多的额度'
     ,historyTCnt:'回复数'
     ,historyTCntInfo:'(max_tokens)回复数越大 ,越有可能消耗更多的额度'
+    ,modelMaxTokens: '模型支持的最大额度'
+    ,modelMaxTokensInfo: '模型支持的最大额度'
     ,role:'角色设定'
     ,rolePlaceholder:'给你的会话设置一个专属的角色，不是必须'
     ,loading2:'正在加载...'
diff --git a/src/locales/zh-TW.ts b/src/locales/zh-TW.ts
index 210066e446..3a53effe4a 100644
--- a/src/locales/zh-TW.ts
+++ b/src/locales/zh-TW.ts
@@ -296,6 +296,8 @@ export default {
     "historyToken": "更多的上下文會使記憶更精確，但會消耗更多的額度",
     "historyTCnt": "回復數",
     "historyTCntInfo": "回復數越大 ,越有可能消耗更多的額度",
+    "modelMaxTokens": "模型支持的最大額度",
+    "modelMaxTokensInfo": "模型支持的最大額度",
     "role": "角色設定",
     "rolePlaceholder": "給你的會話設置一個專屬的角色，不是必須",
     "loading2": "正在加載...",
diff --git a/src/store/homeStore.ts b/src/store/homeStore.ts
index 0cdf01fe9b..21a51e2b9b 100644
--- a/src/store/homeStore.ts
+++ b/src/store/homeStore.ts
@@ -50,6 +50,7 @@ export interface gptConfigType{
     frequency_penalty?:number
     presence_penalty?:number
     tts_voice?:string //TTS 人物
+    model_max_tokens:number // 模型支持最大tokens
 }
 const getGptInt= ():gptConfigType =>{
     let v:gptConfigType=getDefault();
@@ -73,7 +74,8 @@ let v:gptConfigType={
         top_p:1,
         presence_penalty:0,
         frequency_penalty:0,
-        tts_voice:"alloy"
+        tts_voice:"alloy",
+        model_max_tokens:8192
     }
     return v ;
 }
diff --git a/src/views/mj/aiModel.vue b/src/views/mj/aiModel.vue
index 62aeaf0bb0..da1492f10a 100644
--- a/src/views/mj/aiModel.vue
+++ b/src/views/mj/aiModel.vue
@@ -113,7 +113,7 @@ watch(()=>nGptStore.value.model,(n)=>{
          max=4096*2;
     }
 
-    config.value.maxToken=max/2;
+    config.value.maxToken=nGptStore.value.model_max_tokens ?? max/2;
     if(nGptStore.value.max_tokens> config.value.maxToken ) nGptStore.value.max_tokens= config.value.maxToken;
 })
 
@@ -183,12 +183,19 @@ const serverSuccess=(s:any)=>{
     </div>
 </section>
 <div class="mb-4 text-[12px] text-gray-300 dark:text-gray-300/20">{{ $t('mjchat.historyToken') }}</div>
+<section class="flex justify-between items-center">
+  <div>{{ $t('mjchat.modelMaxTokens') }}</div>
+  <div class="w-[80%] max-w-[240px]">
+    <n-input v-model:value="nGptStore.model_max_tokens" />
+  </div>
+</section>
+<div class="mb-4 text-[12px] text-gray-300 dark:text-gray-300/20">{{ $t('mjchat.modelMaxTokensInfo') }}</div>
 
  <section class=" flex justify-between items-center"  >
      <div> {{ $t('mjchat.historyTCnt') }} 
      </div>
      <div class=" flex justify-end items-center w-[80%] max-w-[240px]">
-        <div class=" w-[200px]"><n-slider v-model:value="nGptStore.max_tokens" :step="1" :max="config.maxToken" :min="1" /></div>
+        <div class=" w-[200px]"><n-slider v-model:value="nGptStore.max_tokens" :step="1" :max="nGptStore.model_max_tokens" :min="1" /></div>
         <div  class="w-[50px] text-right">{{ nGptStore.max_tokens }}</div>
     </div>
 </section>