diff --git a/src/main/presenter/configPresenter/modelConfig.ts b/src/main/presenter/configPresenter/modelConfig.ts
index bed26caac..6afeae5c7 100644
--- a/src/main/presenter/configPresenter/modelConfig.ts
+++ b/src/main/presenter/configPresenter/modelConfig.ts
@@ -8,6 +8,8 @@ const SPECIAL_CONCAT_CHAR = '-_-'
 
 export class ModelConfigHelper {
   private modelConfigStore: ElectronStore<Record<string, IModelConfig>>
+  private memoryCache: Map<string, IModelConfig> = new Map()
+  private cacheInitialized: boolean = false
 
   constructor() {
     this.modelConfigStore = new ElectronStore<Record<string, IModelConfig>>({
@@ -15,6 +17,20 @@ export class ModelConfigHelper {
     })
   }
 
+  /**
+   * Initialize memory cache by loading all data from store
+   * This is called lazily on first access
+   */
+  private initializeCache(): void {
+    if (this.cacheInitialized) return
+
+    const allConfigs = this.modelConfigStore.store
+    Object.entries(allConfigs).forEach(([key, value]) => {
+      this.memoryCache.set(key, value)
+    })
+    this.cacheInitialized = true
+  }
+
   /**
    * Get model configuration with priority: user config > provider config > default config
    * @param modelId - The model ID
@@ -22,9 +38,22 @@ export class ModelConfigHelper {
    * @returns ModelConfig
    */
   getModelConfig(modelId: string, providerId?: string): ModelConfig {
+    // Initialize cache if not already done
+    this.initializeCache()
+
     // 1. First try to get user-defined config for this specific provider + model
     if (providerId) {
-      const userConfig = this.modelConfigStore.get(providerId + SPECIAL_CONCAT_CHAR + modelId)
+      const cacheKey = providerId + SPECIAL_CONCAT_CHAR + modelId
+      let userConfig = this.memoryCache.get(cacheKey)
+
+      // If not in cache, try to load from store and cache it
+      if (!userConfig) {
+        userConfig = this.modelConfigStore.get(cacheKey)
+        if (userConfig) {
+          this.memoryCache.set(cacheKey, userConfig)
+        }
+      }
+
       if (userConfig?.config) {
         return userConfig.config
       }
@@ -73,11 +102,16 @@ export class ModelConfigHelper {
    * @param config - The model configuration
    */
   setModelConfig(modelId: string, providerId: string, config: ModelConfig): void {
-    this.modelConfigStore.set(providerId + SPECIAL_CONCAT_CHAR + modelId, {
+    const cacheKey = providerId + SPECIAL_CONCAT_CHAR + modelId
+    const configData: IModelConfig = {
       id: modelId,
       providerId: providerId,
       config: config
-    })
+    }
+
+    // Update both store and cache
+    this.modelConfigStore.set(cacheKey, configData)
+    this.memoryCache.set(cacheKey, configData)
   }
 
   /**
@@ -86,7 +120,11 @@ export class ModelConfigHelper {
    * @param providerId - The provider ID
    */
   resetModelConfig(modelId: string, providerId: string): void {
-    this.modelConfigStore.delete(providerId + SPECIAL_CONCAT_CHAR + modelId)
+    const cacheKey = providerId + SPECIAL_CONCAT_CHAR + modelId
+
+    // Remove from both store and cache
+    this.modelConfigStore.delete(cacheKey)
+    this.memoryCache.delete(cacheKey)
   }
 
   /**
@@ -94,7 +132,15 @@ export class ModelConfigHelper {
    * @returns Record of all configurations
    */
   getAllModelConfigs(): Record<string, IModelConfig> {
-    return this.modelConfigStore.store
+    // Initialize cache if not already done
+    this.initializeCache()
+
+    // Return data from cache for better performance
+    const result: Record<string, IModelConfig> = {}
+    this.memoryCache.forEach((value, key) => {
+      result[key] = value
+    })
+    return result
   }
 
   /**
@@ -126,8 +172,24 @@ export class ModelConfigHelper {
    * @returns boolean
    */
   hasUserConfig(modelId: string, providerId: string): boolean {
-    const userConfig = this.modelConfigStore.get(providerId + SPECIAL_CONCAT_CHAR + modelId)
-    return !!userConfig
+    // Initialize cache if not already done
+    this.initializeCache()
+
+    const cacheKey = providerId + SPECIAL_CONCAT_CHAR + modelId
+
+    // Check cache first
+    if (this.memoryCache.has(cacheKey)) {
+      return true
+    }
+
+    // If not in cache, check store and update cache if found
+    const userConfig = this.modelConfigStore.get(cacheKey)
+    if (userConfig) {
+      this.memoryCache.set(cacheKey, userConfig)
+      return true
+    }
+
+    return false
   }
 
   /**
@@ -137,16 +199,20 @@ export class ModelConfigHelper {
    */
   importConfigs(configs: Record<string, IModelConfig>, overwrite: boolean = false): void {
     if (overwrite) {
-      // Clear existing configs
+      // Clear existing configs from both store and cache
       this.modelConfigStore.clear()
+      this.memoryCache.clear()
     }
 
-    // Import configs
+    // Import configs to both store and cache
     Object.entries(configs).forEach(([key, value]) => {
       if (overwrite || !this.modelConfigStore.has(key)) {
         this.modelConfigStore.set(key, value)
+        this.memoryCache.set(key, value)
       }
     })
+
+    this.cacheInitialized = true
   }
 
   /**
@@ -162,6 +228,7 @@ export class ModelConfigHelper {
    */
   clearAllConfigs(): void {
     this.modelConfigStore.clear()
+    this.memoryCache.clear()
   }
 
   /**
@@ -171,4 +238,12 @@ export class ModelConfigHelper {
   getStorePath(): string {
     return this.modelConfigStore.path
   }
+
+  /**
+   * Clear memory cache (useful for testing or memory management)
+   */
+  clearMemoryCache(): void {
+    this.memoryCache.clear()
+    this.cacheInitialized = false
+  }
 }
diff --git a/src/main/presenter/githubCopilotDeviceFlow.ts b/src/main/presenter/githubCopilotDeviceFlow.ts
index 45ec29dce..30cc9c528 100644
--- a/src/main/presenter/githubCopilotDeviceFlow.ts
+++ b/src/main/presenter/githubCopilotDeviceFlow.ts
@@ -1,9 +1,8 @@
 import { BrowserWindow, shell } from 'electron'
-import { exec } from 'child_process';
+import { exec } from 'child_process'
 import { presenter } from '@/presenter'
 
-
-const GITHUB_DEVICE_URL = 'https://github.com/login/device';
+const GITHUB_DEVICE_URL = 'https://github.com/login/device'
 
 export interface DeviceFlowConfig {
   clientId: string
@@ -303,18 +302,18 @@ export class GitHubCopilotDeviceFlow {
       setTimeout(async () => {
         try {
           // 使用固定的GitHub设备激活页面
-          const url = GITHUB_DEVICE_URL;
-          console.log('Attempting to open URL:', url);
-          
+          const url = GITHUB_DEVICE_URL
+          console.log('Attempting to open URL:', url)
+
           if (process.platform === 'win32') {
             // 先尝试使用explorer命令
             exec(`explorer "${url}"`, (error) => {
               if (error) {
-                console.error('Explorer command failed:', error);
+                console.error('Explorer command failed:', error)
                 // 如果explorer失败，尝试使用start命令
                 exec(`start "" "${url}"`, (startError) => {
                   if (startError) {
-                    console.error('Start command failed:', startError);
+                    console.error('Start command failed:', startError)
                     // 使用更安全的方式处理剪贴板操作
                     instructionWindow.webContents.executeJavaScript(`
                       const shouldCopy = confirm('无法自动打开浏览器。是否复制链接到剪贴板？');
@@ -325,20 +324,20 @@ export class GitHubCopilotDeviceFlow {
                       } else {
                         alert('请手动访问: ${url}');
                       }
-                    `);
+                    `)
                   }
-                });
+                })
               }
-            });
+            })
           } else {
             // 非Windows系统使用默认的shell.openExternal
-            await shell.openExternal(url);
+            await shell.openExternal(url)
           }
         } catch (error) {
-          console.error('Failed to open browser:', error);
+          console.error('Failed to open browser:', error)
           instructionWindow.webContents.executeJavaScript(`
             alert('无法自动打开浏览器，请手动访问: ${GITHUB_DEVICE_URL}');
-          `);
+          `)
         }
       }, 1000)
 
diff --git a/src/main/presenter/llmProviderPresenter/baseProvider.ts b/src/main/presenter/llmProviderPresenter/baseProvider.ts
index 9ab0b8cbf..ed4741ac9 100644
--- a/src/main/presenter/llmProviderPresenter/baseProvider.ts
+++ b/src/main/presenter/llmProviderPresenter/baseProvider.ts
@@ -11,6 +11,8 @@ import {
 import { ConfigPresenter } from '../configPresenter'
 import { DevicePresenter } from '../devicePresenter'
 import { jsonrepair } from 'jsonrepair'
+import { eventBus, SendTarget } from '@/eventbus'
+import { CONFIG_EVENTS } from '@/events'
 
 /**
  * 基础LLM提供商抽象类
@@ -43,6 +45,9 @@ export abstract class BaseLLMProvider {
     this.provider = provider
     this.configPresenter = configPresenter
     this.defaultHeaders = DevicePresenter.getDefaultHeaders()
+
+    // Initialize models and customModels from cached config data
+    this.loadCachedModels()
   }
 
   /**
@@ -53,6 +58,37 @@ export abstract class BaseLLMProvider {
     return BaseLLMProvider.MAX_TOOL_CALLS
   }
 
+  /**
+   * 从配置中加载缓存的模型数据
+   * 在构造函数中调用，避免每次都需要重新获取模型列表
+   */
+  private loadCachedModels(): void {
+    try {
+      // Load cached provider models from config
+      const cachedModels = this.configPresenter.getProviderModels(this.provider.id)
+      if (cachedModels && cachedModels.length > 0) {
+        this.models = cachedModels
+        console.info(
+          `Loaded ${cachedModels.length} cached models for provider: ${this.provider.name}`
+        )
+      }
+
+      // Load cached custom models from config
+      const cachedCustomModels = this.configPresenter.getCustomModels(this.provider.id)
+      if (cachedCustomModels && cachedCustomModels.length > 0) {
+        this.customModels = cachedCustomModels
+        console.info(
+          `Loaded ${cachedCustomModels.length} cached custom models for provider: ${this.provider.name}`
+        )
+      }
+    } catch (error) {
+      console.warn(`Failed to load cached models for provider: ${this.provider.name}`, error)
+      // Keep default empty arrays if loading fails
+      this.models = []
+      this.customModels = []
+    }
+  }
+
   /**
    * 初始化提供商
    * 包括获取模型列表、配置代理等
@@ -79,9 +115,8 @@ export abstract class BaseLLMProvider {
     if (!this.models || this.models.length === 0) return
     const providerId = this.provider.id
 
-    // 检查是否有自定义模型
-    const customModels = this.configPresenter.getCustomModels(providerId)
-    if (customModels && customModels.length > 0) return
+    // 检查是否有自定义模型 (use cached customModels)
+    if (this.customModels && this.customModels.length > 0) return
 
     // 检查是否有任何模型的状态被手动修改过
     const hasManuallyModifiedModels = this.models.some((model) =>
@@ -124,6 +159,22 @@ export abstract class BaseLLMProvider {
     }
   }
 
+  /**
+   * 强制刷新模型数据
+   * 忽略缓存，重新从网络获取最新的模型列表
+   * @returns 模型列表
+   */
+  public async refreshModels(): Promise<void> {
+    console.info(`Force refreshing models for provider: ${this.provider.name}`)
+    await this.fetchModels()
+    await this.autoEnableModelsIfNeeded()
+    eventBus.sendToRenderer(
+      CONFIG_EVENTS.MODEL_LIST_CHANGED,
+      SendTarget.ALL_WINDOWS,
+      this.provider.id
+    )
+  }
+
   /**
    * 获取特定提供商的模型
    * 此方法由具体的提供商子类实现
@@ -160,6 +211,9 @@ export abstract class BaseLLMProvider {
       this.customModels.push(newModel)
     }
 
+    // Sync with config
+    this.configPresenter.addCustomModel(this.provider.id, newModel)
+
     return newModel
   }
 
@@ -172,6 +226,8 @@ export abstract class BaseLLMProvider {
     const index = this.customModels.findIndex((model) => model.id === modelId)
     if (index !== -1) {
       this.customModels.splice(index, 1)
+      // Sync with config
+      this.configPresenter.removeCustomModel(this.provider.id, modelId)
       return true
     }
     return false
@@ -188,6 +244,8 @@ export abstract class BaseLLMProvider {
     if (model) {
       // 应用更新
       Object.assign(model, updates)
+      // Sync with config
+      this.configPresenter.updateCustomModel(this.provider.id, modelId, updates)
       return true
     }
     return false
diff --git a/src/main/presenter/llmProviderPresenter/index.ts b/src/main/presenter/llmProviderPresenter/index.ts
index ac26caad9..5e55624fa 100644
--- a/src/main/presenter/llmProviderPresenter/index.ts
+++ b/src/main/presenter/llmProviderPresenter/index.ts
@@ -1045,6 +1045,17 @@ export class LLMProviderPresenter implements ILlmProviderPresenter {
     return provider.getKeyStatus()
   }
 
+  async refreshModels(providerId: string): Promise<void> {
+    try {
+      const provider = this.getProviderInstance(providerId)
+      await provider.refreshModels()
+    } catch (error) {
+      console.error(`Failed to refresh models for provider ${providerId}:`, error)
+      const errorMessage = error instanceof Error ? error.message : String(error)
+      throw new Error(`Model refresh failed: ${errorMessage}`)
+    }
+  }
+
   async addCustomModel(
     providerId: string,
     model: Omit<MODEL_META, 'providerId' | 'isCustom' | 'group'>
diff --git a/src/main/presenter/llmProviderPresenter/providers/geminiProvider.ts b/src/main/presenter/llmProviderPresenter/providers/geminiProvider.ts
index 128e4053d..32485c86e 100644
--- a/src/main/presenter/llmProviderPresenter/providers/geminiProvider.ts
+++ b/src/main/presenter/llmProviderPresenter/providers/geminiProvider.ts
@@ -24,6 +24,8 @@ import {
 } from '@shared/presenter'
 import { ConfigPresenter } from '../../configPresenter'
 import { BaseLLMProvider, SUMMARY_TITLES_PROMPT } from '../baseProvider'
+import { eventBus, SendTarget } from '@/eventbus'
+import { CONFIG_EVENTS } from '@/events'
 
 // Mapping from simple keys to API HarmCategory constants
 const keyToHarmCategoryMap: Record<string, HarmCategory> = {
@@ -295,6 +297,12 @@ export class GeminiProvider extends BaseLLMProvider {
         // 使用API获取模型列表，如果失败则回退到静态列表
         this.models = await this.fetchProviderModels()
         await this.autoEnableModelsIfNeeded()
+        // gemini 比较慢，特殊补偿一下
+        eventBus.sendToRenderer(
+          CONFIG_EVENTS.MODEL_LIST_CHANGED,
+          SendTarget.ALL_WINDOWS,
+          this.provider.id
+        )
         console.info('Provider initialized successfully:', this.provider.name)
       } catch (error) {
         console.warn('Provider initialization failed:', this.provider.name, error)
diff --git a/src/renderer/src/components/settings/GeminiSafetyConfig.vue b/src/renderer/src/components/settings/GeminiSafetyConfig.vue
index fc6d934a5..79fb4ca44 100644
--- a/src/renderer/src/components/settings/GeminiSafetyConfig.vue
+++ b/src/renderer/src/components/settings/GeminiSafetyConfig.vue
@@ -13,12 +13,12 @@
                   t(setting.label, key.charAt(0).toUpperCase() + key.slice(1))
                 }}</Label>
                 <span class="text-sm text-muted-foreground">{{
-                  t(levelLabels[safetyLevels[key]], `${levelToValueMap[safetyLevels[key]]}`)
+                  t(getLevelLabel(safetyLevels[key]), getLevelValue(safetyLevels[key]))
                 }}</span>
               </div>
               <Slider
                 :id="`${provider.id}-safety-${key}`"
-                :model-value="[safetyLevels[key]]"
+                :model-value="[getSafetyLevel(key)]"
                 :min="0"
                 :max="3"
                 :step="1"
@@ -39,25 +39,24 @@
 </template>
 
 <script setup lang="ts">
-import { reactive, watch } from 'vue'
-import { useI18n } from 'vue-i18n'
-import { Label } from '@/components/ui/label'
-import { Slider } from '@/components/ui/slider'
 import {
   Accordion,
-  AccordionItem,
   AccordionContent,
+  AccordionItem,
   AccordionTrigger
 } from '@/components/ui/accordion'
-import type { LLM_PROVIDER } from '@shared/presenter'
+import { Label } from '@/components/ui/label'
+import { Slider } from '@/components/ui/slider'
 import {
+  levelLabels,
+  levelToValueMap,
   safetyCategories,
   SafetyCategoryKey,
-  SafetySettingValue,
-  levelToValueMap,
-  levelLabels
+  SafetySettingValue
 } from '@/lib/gemini'
-
+import type { LLM_PROVIDER } from '@shared/presenter'
+import { reactive, watch } from 'vue'
+import { useI18n } from 'vue-i18n'
 const { t } = useI18n()
 
 const props = defineProps<{
@@ -71,11 +70,34 @@ const emit = defineEmits<{
 
 const safetyLevels = reactive<Record<string, number>>({})
 
-// 初始化安全级别
+// 安全访问函数
+const getSafetyLevel = (key: string): number => {
+  return safetyLevels[key] ?? safetyCategories[key as SafetyCategoryKey]?.defaultLevel ?? 0
+}
+
+const getLevelLabel = (level: number | undefined): string => {
+  const safeLevel = level ?? 0
+  return levelLabels[safeLevel] ?? levelLabels[0]
+}
+
+const getLevelValue = (level: number | undefined): string => {
+  const safeLevel = level ?? 0
+  return levelToValueMap[safeLevel] ?? levelToValueMap[0]
+}
+
+// 初始化安全级别 - 改进 watch 逻辑
 watch(
   () => props.initialSafetyLevels,
   (newLevels) => {
-    if (newLevels) {
+    // console.log('Safety levels update:', JSON.stringify(safetyLevels), JSON.stringify(newLevels))
+
+    // 清空现有数据
+    Object.keys(safetyLevels).forEach((key) => {
+      delete safetyLevels[key]
+    })
+
+    if (newLevels && Object.keys(newLevels).length > 0) {
+      // 使用新的安全级别
       Object.assign(safetyLevels, newLevels)
     } else {
       // 使用默认值
@@ -83,8 +105,10 @@ watch(
         safetyLevels[key] = safetyCategories[key as SafetyCategoryKey].defaultLevel
       }
     }
+
+    // console.log('Final safety levels:', JSON.stringify(safetyLevels))
   },
-  { immediate: true }
+  { immediate: true, deep: true }
 )
 
 const handleSafetySettingChange = (key: SafetyCategoryKey, level: number) => {
diff --git a/src/renderer/src/components/settings/ModelProviderSettingsDetail.vue b/src/renderer/src/components/settings/ModelProviderSettingsDetail.vue
index a7ba3199f..0c87001fd 100644
--- a/src/renderer/src/components/settings/ModelProviderSettingsDetail.vue
+++ b/src/renderer/src/components/settings/ModelProviderSettingsDetail.vue
@@ -26,7 +26,7 @@
         <GeminiSafetyConfig
           v-if="provider.id === 'gemini'"
           :provider="provider"
-          :initial-safety-levels="geminiSafetyLevels"
+          :initial-safety-levels="geminiSafetyLevelsForChild"
           @safety-setting-change="handleSafetySettingChange"
         />
 
@@ -75,6 +75,7 @@ import ProviderModelManager from './ProviderModelManager.vue'
 import ProviderDialogContainer from './ProviderDialogContainer.vue'
 import { useModelCheckStore } from '@/stores/modelCheck'
 import { levelToValueMap, safetyCategories } from '@/lib/gemini'
+import type { SafetyCategoryKey, SafetySettingValue } from '@/lib/gemini'
 
 interface ProviderWebsites {
   official: string
@@ -84,15 +85,9 @@ interface ProviderWebsites {
   defaultBaseUrl: string
 }
 
-// Define safety types for Gemini safety handling
-type SafetyCategoryKey = 'harassment' | 'hateSpeech' | 'sexuallyExplicit' | 'dangerousContent'
-type SafetySettingValue =
-  | 'BLOCK_NONE'
-  | 'BLOCK_LOW_AND_ABOVE'
-  | 'BLOCK_MEDIUM_AND_ABOVE'
-  | 'BLOCK_ONLY_HIGH'
-  | 'HARM_BLOCK_THRESHOLD_UNSPECIFIED'
+// Types are imported from @/lib/gemini
 
+// Value to level mapping for Gemini safety settings
 const valueToLevelMap: Record<SafetySettingValue, number> = {
   BLOCK_NONE: 0,
   BLOCK_LOW_AND_ABOVE: 1,
@@ -200,8 +195,13 @@ const initData = async () => {
   // Fetch Gemini Safety Settings if applicable
   if (props.provider.id === 'gemini') {
     console.log('Fetching Gemini safety settings...')
+
+    // 先清空现有数据
+    Object.keys(geminiSafetyLevels).forEach((key) => {
+      delete geminiSafetyLevels[key]
+    })
+
     for (const key in safetyCategories) {
-      console.log('key:', key)
       const categoryKey = key as string
       try {
         const savedValue = (await settingsStore.getGeminiSafety(categoryKey)) as
@@ -209,13 +209,17 @@ const initData = async () => {
           | 'HARM_BLOCK_THRESHOLD_UNSPECIFIED'
         console.log(`Fetched Gemini safety for ${categoryKey}:`, savedValue)
         geminiSafetyLevels[categoryKey] =
-          valueToLevelMap[savedValue] ?? safetyCategories[categoryKey].defaultLevel
+          valueToLevelMap[savedValue as SafetySettingValue] ??
+          safetyCategories[categoryKey as SafetyCategoryKey].defaultLevel
         console.log(`Set Gemini level for ${categoryKey}:`, geminiSafetyLevels[categoryKey])
       } catch (error) {
         console.error(`Failed to fetch Gemini safety setting for ${categoryKey}:`, error)
-        geminiSafetyLevels[categoryKey] = safetyCategories[categoryKey].defaultLevel // Default on error
+        geminiSafetyLevels[categoryKey] =
+          safetyCategories[categoryKey as SafetyCategoryKey].defaultLevel // Default on error
       }
     }
+
+    console.log('All Gemini safety levels initialized:', JSON.stringify(geminiSafetyLevels))
   }
 }
 
@@ -348,4 +352,10 @@ const handleConfigChanged = async () => {
 const openModelCheckDialog = () => {
   modelCheckStore.openDialog(props.provider.id)
 }
+
+// 使用 computed 确保响应性正确传递
+const geminiSafetyLevelsForChild = computed(() => {
+  // 创建一个新的对象确保响应性
+  return { ...geminiSafetyLevels }
+})
 </script>
diff --git a/src/renderer/src/components/settings/ProviderApiConfig.vue b/src/renderer/src/components/settings/ProviderApiConfig.vue
index 6909b093e..839932735 100644
--- a/src/renderer/src/components/settings/ProviderApiConfig.vue
+++ b/src/renderer/src/components/settings/ProviderApiConfig.vue
@@ -67,10 +67,17 @@
           variant="outline"
           size="xs"
           class="text-xs text-normal rounded-lg"
-          @click="openProviderWebsite"
+          :disabled="isRefreshing"
+          @click="refreshModels"
         >
-          <Icon icon="lucide:hand-helping" class="w-4 h-4 text-muted-foreground" />{{
-            t('settings.provider.howToGet')
+          <Icon
+            :icon="isRefreshing ? 'lucide:loader-2' : 'lucide:refresh-cw'"
+            :class="['w-4 h-4 text-muted-foreground', { 'animate-spin': isRefreshing }]"
+          />
+          {{
+            isRefreshing
+              ? t('settings.provider.refreshingModels')
+              : t('settings.provider.refreshModels')
           }}
         </Button>
         <!-- Key Status Display -->
@@ -94,7 +101,7 @@
         </div>
       </div>
       <div v-if="!provider.custom" class="text-xs text-muted-foreground">
-        {{ t('settings.provider.getKeyTip') }}
+        {{ t('settings.provider.howToGet') }}: {{ t('settings.provider.getKeyTip') }}
         <a :href="providerWebsites?.apiKey" target="_blank" class="text-primary">{{
           provider.name
         }}</a>
@@ -145,6 +152,7 @@ const emit = defineEmits<{
 const apiKey = ref(props.provider.apiKey || '')
 const apiHost = ref(props.provider.baseUrl || '')
 const keyStatus = ref<KeyStatus | null>(null)
+const isRefreshing = ref(false)
 
 watch(
   () => props.provider,
@@ -163,13 +171,6 @@ const handleApiHostChange = (value: string) => {
   emit('api-host-change', value)
 }
 
-const openProviderWebsite = () => {
-  const url = props.providerWebsites?.apiKey
-  if (url) {
-    window.open(url, '_blank')
-  }
-}
-
 const handleOAuthSuccess = () => {
   emit('oauth-success')
 }
@@ -198,6 +199,19 @@ const getKeyStatus = async () => {
   }
 }
 
+const refreshModels = async () => {
+  if (isRefreshing.value) return
+
+  isRefreshing.value = true
+  try {
+    await llmProviderPresenter.refreshModels(props.provider.id)
+  } catch (error) {
+    console.error('Failed to refresh models:', error)
+  } finally {
+    isRefreshing.value = false
+  }
+}
+
 onMounted(() => {
   getKeyStatus()
 })
diff --git a/src/renderer/src/i18n/en-US/settings.json b/src/renderer/src/i18n/en-US/settings.json
index 233c0ece4..90959bdd3 100644
--- a/src/renderer/src/i18n/en-US/settings.json
+++ b/src/renderer/src/i18n/en-US/settings.json
@@ -248,7 +248,11 @@
     "localModels": "Local Models",
     "azureApiVersion": "API Version",
     "safety": {
-      "title": "Safety settings"
+      "title": "Safety settings",
+      "blockHighest": "Block high risk",
+      "blockMost": "Block medium risks",
+      "blockNone": "Not blocked",
+      "blockSome": "Block low risk"
     },
     "serverList": "Server List",
     "totalServers": "Total Servers",
@@ -270,7 +274,8 @@
     "keyStatus": {
       "remaining": "Remaining amount",
       "usage": "Used"
-    }
+    },
+    "refreshingModels": "Refreshing..."
   },
   "knowledgeBase": {
     "title": "Knowledge Base Settings",
diff --git a/src/renderer/src/i18n/fa-IR/settings.json b/src/renderer/src/i18n/fa-IR/settings.json
index 662cde6d3..58bbb4023 100644
--- a/src/renderer/src/i18n/fa-IR/settings.json
+++ b/src/renderer/src/i18n/fa-IR/settings.json
@@ -248,7 +248,11 @@
     "localModels": "مدل‌های محلی",
     "azureApiVersion": "نگارش API",
     "safety": {
-      "title": "تنظیمات ایمنی"
+      "title": "تنظیمات ایمنی",
+      "blockHighest": "ریسک بالا",
+      "blockMost": "خطرات متوسط ​​را مسدود کنید",
+      "blockNone": "مسدود نشده",
+      "blockSome": "بلوک ریسک کم"
     },
     "serverList": "فهرست کارسازها",
     "totalServers": "مجموع کارسازها",
@@ -270,7 +274,8 @@
     "keyStatus": {
       "remaining": "مقدار باقی مانده",
       "usage": "مورد استفاده"
-    }
+    },
+    "refreshingModels": "طراوت ..."
   },
   "knowledgeBase": {
     "title": "تنظیمات پایگاه دانش",
diff --git a/src/renderer/src/i18n/fr-FR/settings.json b/src/renderer/src/i18n/fr-FR/settings.json
index 212e11a82..efb74d64a 100644
--- a/src/renderer/src/i18n/fr-FR/settings.json
+++ b/src/renderer/src/i18n/fr-FR/settings.json
@@ -248,7 +248,11 @@
     "localModels": "Modèles locaux",
     "azureApiVersion": "Version API",
     "safety": {
-      "title": "Paramètres de sécurité"
+      "title": "Paramètres de sécurité",
+      "blockHighest": "Bloquer un risque élevé",
+      "blockMost": "Bloquer les risques moyens",
+      "blockNone": "Non bloqué",
+      "blockSome": "Bloquer un faible risque"
     },
     "serverList": "Liste des serveurs",
     "totalServers": "Total des serveurs",
@@ -270,7 +274,8 @@
     "keyStatus": {
       "remaining": "Montant restant",
       "usage": "Utilisé"
-    }
+    },
+    "refreshingModels": "Rafraîchissant..."
   },
   "knowledgeBase": {
     "title": "Paramètres de la base de connaissances",
diff --git a/src/renderer/src/i18n/ja-JP/settings.json b/src/renderer/src/i18n/ja-JP/settings.json
index 3b15f62b6..8c15b0936 100644
--- a/src/renderer/src/i18n/ja-JP/settings.json
+++ b/src/renderer/src/i18n/ja-JP/settings.json
@@ -248,7 +248,11 @@
     "localModels": "ローカルモデル",
     "azureApiVersion": "APIバージョン",
     "safety": {
-      "title": "セキュリティ設定"
+      "title": "セキュリティ設定",
+      "blockHighest": "高リスクをブロックします",
+      "blockMost": "中程度のリスクをブロックします",
+      "blockNone": "ブロックされていません",
+      "blockSome": "低リスクをブロックします"
     },
     "serverList": "サーバーリスト",
     "totalServers": "サーバー総数",
@@ -270,7 +274,8 @@
     "keyStatus": {
       "remaining": "残りの金額",
       "usage": "使用済み"
-    }
+    },
+    "refreshingModels": "更新中..."
   },
   "knowledgeBase": {
     "title": "ナレッジベース設定",
diff --git a/src/renderer/src/i18n/ko-KR/settings.json b/src/renderer/src/i18n/ko-KR/settings.json
index dcb9b7877..2eaa43337 100644
--- a/src/renderer/src/i18n/ko-KR/settings.json
+++ b/src/renderer/src/i18n/ko-KR/settings.json
@@ -248,7 +248,11 @@
     "urlFormat": "API 샘플 : {defaultUrl}",
     "azureApiVersion": "API 버전",
     "safety": {
-      "title": "보안 설정"
+      "title": "보안 설정",
+      "blockHighest": "높은 위험을 차단하십시오",
+      "blockMost": "중간 위험을 차단합니다",
+      "blockNone": "차단되지 않았습니다",
+      "blockSome": "낮은 위험을 차단하십시오"
     },
     "serverList": "서버 목록",
     "totalServers": "총 서버 수",
@@ -270,7 +274,8 @@
     "keyStatus": {
       "remaining": "나머지 금액",
       "usage": "사용된"
-    }
+    },
+    "refreshingModels": "상쾌한 ..."
   },
   "knowledgeBase": {
     "title": "지식 베이스 설정",
diff --git a/src/renderer/src/i18n/ru-RU/settings.json b/src/renderer/src/i18n/ru-RU/settings.json
index 98e08acd2..fb75db93f 100644
--- a/src/renderer/src/i18n/ru-RU/settings.json
+++ b/src/renderer/src/i18n/ru-RU/settings.json
@@ -248,7 +248,11 @@
     "urlFormat": "Образец API: {defaultUrl}",
     "azureApiVersion": "Версия API",
     "safety": {
-      "title": "Настройки безопасности"
+      "title": "Настройки безопасности",
+      "blockHighest": "Блок высокий риск",
+      "blockMost": "Блок средних рисков",
+      "blockNone": "Не заблокирован",
+      "blockSome": "Блок низкий риск"
     },
     "serverList": "Список серверов",
     "totalServers": "Всего серверов",
@@ -270,7 +274,8 @@
     "keyStatus": {
       "remaining": "Оставшаяся сумма",
       "usage": "Использовал"
-    }
+    },
+    "refreshingModels": "Освежает ..."
   },
   "knowledgeBase": {
     "title": "Настройки базы знаний",
diff --git a/src/renderer/src/i18n/zh-CN/settings.json b/src/renderer/src/i18n/zh-CN/settings.json
index f6b54c62f..adee1c25d 100644
--- a/src/renderer/src/i18n/zh-CN/settings.json
+++ b/src/renderer/src/i18n/zh-CN/settings.json
@@ -165,6 +165,7 @@
     "keyPlaceholder": "请输入API Key",
     "verifyKey": "验证密钥",
     "howToGet": "如何获取",
+    "refreshingModels": "刷新中...",
     "getKeyTip": "请前往",
     "getKeyTipEnd": "获取API Key",
     "urlFormat": "API样例：{defaultUrl}",
@@ -248,7 +249,11 @@
     "localModels": "本地模型",
     "azureApiVersion": "API 版本",
     "safety": {
-      "title": "安全设置"
+      "title": "安全设置",
+      "blockNone": "不屏蔽",
+      "blockSome": "屏蔽低风险",
+      "blockMost": "屏蔽中风险",
+      "blockHighest": "屏蔽高风险"
     },
     "serverList": "服务器列表",
     "totalServers": "服务器总数",
diff --git a/src/renderer/src/i18n/zh-HK/settings.json b/src/renderer/src/i18n/zh-HK/settings.json
index 138fd970f..24c77146d 100644
--- a/src/renderer/src/i18n/zh-HK/settings.json
+++ b/src/renderer/src/i18n/zh-HK/settings.json
@@ -248,7 +248,11 @@
     "modelsEnabled": "模型已經啟用",
     "azureApiVersion": "API 版本",
     "safety": {
-      "title": "安全設置"
+      "title": "安全設置",
+      "blockHighest": "屏蔽高風險",
+      "blockMost": "屏蔽中風險",
+      "blockNone": "不屏蔽",
+      "blockSome": "屏蔽低風險"
     },
     "serverList": "服務器列表",
     "totalServers": "服務器總數",
@@ -270,7 +274,8 @@
     "keyStatus": {
       "remaining": "剩餘額度",
       "usage": "已使用"
-    }
+    },
+    "refreshingModels": "刷新中..."
   },
   "knowledgeBase": {
     "fastgptTitle": "FastGPT知識庫",
diff --git a/src/renderer/src/i18n/zh-TW/settings.json b/src/renderer/src/i18n/zh-TW/settings.json
index 49b92c9d1..9d7aafe70 100644
--- a/src/renderer/src/i18n/zh-TW/settings.json
+++ b/src/renderer/src/i18n/zh-TW/settings.json
@@ -248,7 +248,11 @@
     "noLocalModels": "沒有本機模型",
     "azureApiVersion": "API 版本",
     "safety": {
-      "title": "安全設置"
+      "title": "安全設置",
+      "blockHighest": "屏蔽高風險",
+      "blockMost": "屏蔽中風險",
+      "blockNone": "不屏蔽",
+      "blockSome": "屏蔽低風險"
     },
     "serverList": "伺服器列表",
     "totalServers": "伺服器總數",
@@ -270,7 +274,8 @@
     "keyStatus": {
       "remaining": "剩餘額度",
       "usage": "已使用"
-    }
+    },
+    "refreshingModels": "刷新中..."
   },
   "knowledgeBase": {
     "title": "知識庫設置",
diff --git a/src/renderer/src/stores/settings.ts b/src/renderer/src/stores/settings.ts
index 0aa15d765..47e8a3b21 100644
--- a/src/renderer/src/stores/settings.ts
+++ b/src/renderer/src/stores/settings.ts
@@ -532,14 +532,14 @@ export const useSettingsStore = defineStore('settings', () => {
 
     try {
       // 自定义模型直接从配置存储获取，不需要等待provider实例
-      await refreshCustomModels(providerId)
+      refreshCustomModels(providerId)
 
       // 标准模型需要provider实例，可能需要等待实例初始化
-      await refreshStandardModels(providerId)
+      refreshStandardModels(providerId)
     } catch (error) {
       console.error(`刷新模型失败: ${providerId}`, error)
       // 如果标准模型刷新失败，至少确保自定义模型可用
-      await refreshCustomModels(providerId)
+      refreshCustomModels(providerId)
     }
   }
 
@@ -734,7 +734,7 @@ export const useSettingsStore = defineStore('settings', () => {
       const success = await llmP.removeCustomModel(providerId, modelId)
       console.log('removeCustomModel', providerId, modelId, success)
       if (success) {
-        await refreshCustomModels(providerId) // 只刷新自定义模型
+        refreshCustomModels(providerId) // 只刷新自定义模型
       }
       return success
     } catch (error) {
@@ -753,7 +753,7 @@ export const useSettingsStore = defineStore('settings', () => {
       // 不包含启用状态的常规更新
       const success = await llmP.updateCustomModel(providerId, modelId, updates)
       if (success) {
-        await refreshCustomModels(providerId) // 只刷新自定义模型
+        refreshCustomModels(providerId) // 只刷新自定义模型
       }
       return success
     } catch (error) {
@@ -770,7 +770,7 @@ export const useSettingsStore = defineStore('settings', () => {
     try {
       const newModel = await llmP.addCustomModel(providerId, model)
       await configP.addCustomModel(providerId, newModel)
-      await refreshCustomModels(providerId) // 只刷新自定义模型
+      refreshCustomModels(providerId) // 只刷新自定义模型
       return newModel
     } catch (error) {
       console.error('Failed to add custom model:', error)
diff --git a/src/shared/presenter.d.ts b/src/shared/presenter.d.ts
index d12bd6c2d..c09548b4c 100644
--- a/src/shared/presenter.d.ts
+++ b/src/shared/presenter.d.ts
@@ -514,6 +514,7 @@ export interface ILlmProviderPresenter {
   stopStream(eventId: string): Promise<void>
   check(providerId: string, modelId?: string): Promise<{ isOk: boolean; errorMsg: string | null }>
   getKeyStatus(providerId: string): Promise<KeyStatus | null>
+  refreshModels(providerId: string): Promise<void>
   summaryTitles(
     messages: { role: 'system' | 'user' | 'assistant'; content: string }[],
     providerId: string,