fix: resolve OpenRouter model sync timeout and welcome view UI issues

src/main/presenter/llmProviderPresenter/baseProvider.ts

@@ -30,6 +30,7 @@ import { CONFIG_EVENTS } from '@/events'
 export abstract class BaseLLMProvider {
   // 单轮会话中最大工具调用次数限制
   protected static readonly MAX_TOOL_CALLS = 50
+  protected static readonly DEFAULT_MODEL_FETCH_TIMEOUT = 12000 // 提升到12秒作为通用默认值
 
   protected provider: LLM_PROVIDER
   protected models: MODEL_META[] = []
@@ -59,6 +60,14 @@ export abstract class BaseLLMProvider {
     return BaseLLMProvider.MAX_TOOL_CALLS
   }
 
+  /**
+   * 获取模型获取超时时间配置
+   * @returns 超时时间（毫秒）
+   */
+  protected getModelFetchTimeout(): number {
+    return BaseLLMProvider.DEFAULT_MODEL_FETCH_TIMEOUT
+  }
+
   /**
    * 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

src/main/presenter/llmProviderPresenter/providers/openAICompatibleProvider.ts

@@ -1198,8 +1198,8 @@ export class OpenAICompatibleProvider extends BaseLLMProvider {
   public async check(): Promise<{ isOk: boolean; errorMsg: string | null }> {
     try {
       if (!this.isNoModelsApi) {
-        // Use a reasonable timeout
-        const models = await this.fetchOpenAIModels({ timeout: 5000 }) // Increased timeout slightly
+        // Use unified timeout configuration from base class
+        const models = await this.fetchOpenAIModels({ timeout: this.getModelFetchTimeout() })
         this.models = models // Store fetched models
       }
       // Potentially add a simple API call test here if needed, e.g., list models even for no-API list to check key/endpoint

src/main/presenter/llmProviderPresenter/providers/openAIResponsesProvider.ts

@@ -1070,8 +1070,8 @@ export class OpenAIResponsesProvider extends BaseLLMProvider {
   public async check(): Promise<{ isOk: boolean; errorMsg: string | null }> {
     try {
       if (!this.isNoModelsApi) {
-        // Use a reasonable timeout
-        const models = await this.fetchOpenAIModels({ timeout: 5000 }) // Increased timeout slightly
+        // Use unified timeout configuration from base class
+        const models = await this.fetchOpenAIModels({ timeout: this.getModelFetchTimeout() })
         this.models = models // Store fetched models
       }
       // Potentially add a simple API call test here if needed, e.g., list models even for no-API list to check key/endpoint

src/renderer/src/views/WelcomeView.vue

@@ -211,9 +211,9 @@ onMounted(() => {
 
 const handleModelEnabledChange = async (model: MODEL_META, enabled: boolean) => {
   try {
-    await settingsStore.updateModelStatus(selectedProvider.value, model.id, !enabled)
+    await settingsStore.updateModelStatus(selectedProvider.value, model.id, enabled)
   } catch (error) {
-    console.error('Failed to disable model:', error)
+    console.error('Failed to update model status:', error)
   }
   console.log('handleModelEnabledChange', model, enabled)
 }
@@ -339,7 +339,7 @@ const isFirstStep = computed(() => currentStep.value === 0)
 
               <div
                 v-show="!providerModelLoading"
-                class="flex flex-col w-full border overflow-hidden rounded-lg max-h-80 overflow-y-auto"
+                class="flex flex-col w-full border rounded-lg max-h-80 overflow-y-auto"
               >
                 <ModelConfigItem
                   v-for="model in settingsStore.allProviderModels.find(