@lobehub/chat 1.110.1 → 1.110.2

package/locales/zh-CN/modelProvider.json

@@ -306,6 +306,7 @@
         "latestTime": "上次更新时间：{{time}}",
         "noLatestTime": "暂未获取列表"
       },
+      "noModelsInCategory": "该分类下暂无启用的模型",
       "resetAll": {
         "conform": "确认重置当前模型的所有修改？重置后当前模型列表将会回到默认状态",
         "success": "重置成功",
@@ -316,7 +317,15 @@
       "title": "模型列表",
       "total": "共 {{count}} 个模型可用"
     },
-    "searchNotFound": "未找到搜索结果"
+    "searchNotFound": "未找到搜索结果",
+    "tabs": {
+      "all": "全部",
+      "chat": "对话",
+      "embedding": "向量化",
+      "image": "图片",
+      "stt": "ASR",
+      "tts": "TTS"
+    }
   },
   "sortModal": {
     "success": "排序更新成功",

package/locales/zh-CN/models.json

@@ -74,6 +74,9 @@
   "DeepSeek-V3": {
     "description": "DeepSeek-V3 是一款由深度求索公司自研的MoE模型。DeepSeek-V3 多项评测成绩超越了 Qwen2.5-72B 和 Llama-3.1-405B 等其他开源模型，并在性能上和世界顶尖的闭源模型 GPT-4o 以及 Claude-3.5-Sonnet 不分伯仲。"
   },
+  "DeepSeek-V3-Fast": {
+    "description": "模型供应商为：sophnet平台。DeepSeek V3 Fast 是 DeepSeek V3 0324 版本的高TPS极速版，满血非量化，代码与数学能力更强，响应更快！"
+  },
   "Doubao-lite-128k": {
     "description": "Doubao-lite 拥有极致的响应速度，更好的性价比，为客户不同场景提供更灵活的选择。支持128k上下文窗口的推理和精调。"
   },
@@ -608,6 +611,9 @@
   "aya:35b": {
     "description": "Aya 23 是 Cohere 推出的多语言模型，支持 23 种语言，为多元化语言应用提供便利。"
   },
+  "azure-DeepSeek-R1-0528": {
+    "description": "由微软部署提供； DeepSeek R1型号已进行小版本升级，当前版本为DeepSeek-R1-0528。在最新的更新中，DeepSeek R1通过增加计算资源和引入后训练阶段的算法优化机制，大幅提升了推理深度和推断能力。该模型在数学、编程和通用逻辑等多个基准测试中表现出色，其整体性能已接近领先模型，如O3和Gemini 2.5 Pro 。"
+  },
   "baichuan/baichuan2-13b-chat": {
     "description": "Baichuan-13B 百川智能开发的包含 130 亿参数的开源可商用的大规模语言模型，在权威的中文和英文 benchmark 上均取得同尺寸最好的效果"
   },
@@ -668,6 +674,9 @@
   "claude-3-sonnet-20240229": {
     "description": "Claude 3 Sonnet 在智能和速度方面为企业工作负载提供了理想的平衡。它以更低的价格提供最大效用，可靠且适合大规模部署。"
   },
+  "claude-opus-4-1-20250805": {
+    "description": "Claude Opus 4.1 是 Anthropic 最新的用于处理高度复杂任务的最强大模型。它在性能、智能、流畅性和理解力方面表现卓越。"
+  },
   "claude-opus-4-20250514": {
     "description": "Claude Opus 4 是 Anthropic 用于处理高度复杂任务的最强大模型。它在性能、智能、流畅性和理解力方面表现卓越。"
   },
@@ -2123,6 +2132,12 @@
   "openai/gpt-4o-mini": {
     "description": "GPT-4o mini是OpenAI在GPT-4 Omni之后推出的最新模型，支持图文输入并输出文本。作为他们最先进的小型模型，它比其他近期的前沿模型便宜很多，并且比GPT-3.5 Turbo便宜超过60%。它保持了最先进的智能，同时具有显著的性价比。GPT-4o mini在MMLU测试中获得了 82% 的得分，目前在聊天偏好上排名高于 GPT-4。"
   },
+  "openai/gpt-oss-120b": {
+    "description": "OpenAI GPT-OSS 120B 是一款拥有 1200 亿参数的顶尖语言模型，内置浏览器搜索和代码执行功能，并具备推理能力。"
+  },
+  "openai/gpt-oss-20b": {
+    "description": "OpenAI GPT-OSS 20B 是一款拥有 200 亿参数的顶尖语言模型，内置浏览器搜索和代码执行功能，并具备推理能力。"
+  },
   "openai/o1": {
     "description": "o1是OpenAI新的推理模型，支持图文输入并输出文本，适用于需要广泛通用知识的复杂任务。该模型具有200K上下文和2023年10月的知识截止日期。"
   },
@@ -2411,8 +2426,11 @@
   "qwen3-coder-480b-a35b-instruct": {
     "description": "通义千问代码模型开源版。最新的 qwen3-coder-480b-a35b-instruct 是基于 Qwen3 的代码生成模型，具有强大的Coding Agent能力，擅长工具调用和环境交互，能够实现自主编程、代码能力卓越的同时兼具通用能力。"
   },
+  "qwen3-coder-flash": {
+    "description": "通义千问代码模型。最新的 Qwen3-Coder 系列模型是基于 Qwen3 的代码生成模型，具有强大的Coding Agent能力，擅长工具调用和环境交互，能够实现自主编程，代码能力卓越的同时兼具通用能力。"
+  },
   "qwen3-coder-plus": {
-    "description": "通义千问代码模型。最新的 Qwen3-Coder-Plus 系列模型是基于 Qwen3 的代码生成模型，具有强大的Coding Agent能力，擅长工具调用和环境交互，能够实现自主编程，代码能力卓越的同时兼具通用能力。"
+    "description": "通义千问代码模型。最新的 Qwen3-Coder 系列模型是基于 Qwen3 的代码生成模型，具有强大的Coding Agent能力，擅长工具调用和环境交互，能够实现自主编程，代码能力卓越的同时兼具通用能力。"
   },
   "qwq": {
     "description": "QwQ 是 Qwen 系列的推理模型。与传统的指令调优模型相比，QwQ 具备思考和推理的能力，能够在下游任务中，尤其是困难问题上，显著提升性能。QwQ-32B 是中型推理模型，能够在与最先进的推理模型（如 DeepSeek-R1、o1-mini）竞争时取得可观的表现。"

package/locales/zh-TW/modelProvider.json

@@ -306,6 +306,7 @@
         "latestTime": "上次更新時間：{{time}}",
         "noLatestTime": "尚未取得列表"
       },
+      "noModelsInCategory": "該分類下暫無啟用的模型",
       "resetAll": {
         "conform": "確認重置當前模型的所有修改？重置後當前模型列表將會回到預設狀態",
         "success": "重置成功",
@@ -316,7 +317,15 @@
       "title": "模型列表",
       "total": "共 {{count}} 個模型可用"
     },
-    "searchNotFound": "未找到搜尋結果"
+    "searchNotFound": "未找到搜尋結果",
+    "tabs": {
+      "all": "全部",
+      "chat": "對話",
+      "embedding": "向量化",
+      "image": "圖片",
+      "stt": "ASR",
+      "tts": "TTS"
+    }
   },
   "sortModal": {
     "success": "排序更新成功",

package/locales/zh-TW/models.json

@@ -74,6 +74,9 @@
   "DeepSeek-V3": {
     "description": "DeepSeek-V3 是一款由深度求索公司自研的MoE模型。DeepSeek-V3 多項評測成績超越了 Qwen2.5-72B 和 Llama-3.1-405B 等其他開源模型，並在性能上和世界頂尖的閉源模型 GPT-4o 以及 Claude-3.5-Sonnet 不分伯仲。"
   },
+  "DeepSeek-V3-Fast": {
+    "description": "模型供應商為：sophnet平台。DeepSeek V3 Fast 是 DeepSeek V3 0324 版本的高TPS極速版，滿血非量化，代碼與數學能力更強，響應更快！"
+  },
   "Doubao-lite-128k": {
     "description": "Doubao-lite 擁有極致的響應速度，更好的性價比，為客戶不同場景提供更靈活的選擇。支持128k上下文視窗的推理和精調。"
   },
@@ -608,6 +611,9 @@
   "aya:35b": {
     "description": "Aya 23 是 Cohere 推出的多語言模型，支持 23 種語言，為多元化語言應用提供便利。"
   },
+  "azure-DeepSeek-R1-0528": {
+    "description": "由微軟部署提供； DeepSeek R1型號已進行小版本升級，當前版本為DeepSeek-R1-0528。在最新的更新中，DeepSeek R1透過增加計算資源和引入後訓練階段的演算法優化機制，大幅提升了推理深度和推斷能力。該模型在數學、程式設計和通用邏輯等多個基準測試中表現出色，其整體性能已接近領先模型，如O3和Gemini 2.5 Pro 。"
+  },
   "baichuan/baichuan2-13b-chat": {
     "description": "Baichuan-13B百川智能開發的包含130億參數的開源可商用的大規模語言模型，在權威的中文和英文benchmark上均取得同尺寸最好的效果。"
   },
@@ -668,6 +674,9 @@
   "claude-3-sonnet-20240229": {
     "description": "Claude 3 Sonnet 在智能和速度方面為企業工作負載提供了理想的平衡。它以更低的價格提供最大效用，可靠且適合大規模部署。"
   },
+  "claude-opus-4-1-20250805": {
+    "description": "Claude Opus 4.1 是 Anthropic 最新的用於處理高度複雜任務的最強大模型。它在性能、智慧、流暢性和理解力方面表現卓越。"
+  },
   "claude-opus-4-20250514": {
     "description": "Claude Opus 4 是 Anthropic 用於處理高度複雜任務的最強大模型。它在性能、智能、流暢性和理解力方面表現卓越。"
   },
@@ -2123,6 +2132,12 @@
   "openai/gpt-4o-mini": {
     "description": "GPT-4o mini是OpenAI在GPT-4 Omni之後推出的最新模型，支持圖文輸入並輸出文本。作為他們最先進的小型模型，它比其他近期的前沿模型便宜很多，並且比GPT-3.5 Turbo便宜超過60%。它保持了最先進的智能，同時具有顯著的性價比。GPT-4o mini在MMLU測試中獲得了82%的得分，目前在聊天偏好上排名高於GPT-4。"
   },
+  "openai/gpt-oss-120b": {
+    "description": "OpenAI GPT-OSS 120B 是一款擁有 1200 億參數的頂尖語言模型，內建瀏覽器搜尋和程式碼執行功能，並具備推理能力。"
+  },
+  "openai/gpt-oss-20b": {
+    "description": "OpenAI GPT-OSS 20B 是一款擁有 200 億參數的頂尖語言模型，內建瀏覽器搜尋和程式碼執行功能，並具備推理能力。"
+  },
   "openai/o1": {
     "description": "o1 是 OpenAI 新的推理模型，支援圖文輸入並輸出文本，適用於需要廣泛通用知識的複雜任務。該模型具有 200K 上下文和 2023 年 10 月的知識截止日期。"
   },
@@ -2411,8 +2426,11 @@
   "qwen3-coder-480b-a35b-instruct": {
     "description": "通義千問程式碼模型開源版。最新的 qwen3-coder-480b-a35b-instruct 是基於 Qwen3 的程式碼生成模型，具有強大的 Coding Agent 能力，擅長工具調用和環境互動，能夠實現自主程式設計、程式碼能力卓越的同時兼具通用能力。"
   },
+  "qwen3-coder-flash": {
+    "description": "通義千問程式碼模型。最新的 Qwen3-Coder 系列模型是基於 Qwen3 的程式碼生成模型，具有強大的Coding Agent能力，擅長工具調用和環境互動，能夠實現自主程式設計，程式碼能力卓越的同時兼具通用能力。"
+  },
   "qwen3-coder-plus": {
-    "description": "通義千問程式碼模型。最新的 Qwen3-Coder-Plus 系列模型是基於 Qwen3 的程式碼生成模型，具有強大的 Coding Agent 能力，擅長工具調用和環境互動，能夠實現自主程式設計，程式碼能力卓越的同時兼具通用能力。"
+    "description": "通義千問程式碼模型。最新的 Qwen3-Coder 系列模型是基於 Qwen3 的程式碼生成模型，具有強大的Coding Agent能力，擅長工具調用和環境互動，能夠實現自主程式設計，程式碼能力卓越的同時兼具通用能力。"
   },
   "qwq": {
     "description": "QwQ 是一個實驗研究模型，專注於提高 AI 推理能力。"

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@lobehub/chat",
-  "version": "1.110.1",
+  "version": "1.110.2",
   "description": "Lobe Chat - an open-source, high-performance chatbot framework that supports speech synthesis, multimodal, and extensible Function Call plugin system. Supports one-click free deployment of your private ChatGPT/LLM web application.",
   "keywords": [
     "framework",

package/packages/electron-client-ipc/src/events/menu.ts

@@ -1,5 +1,5 @@
 export interface MenuDispatchEvents {
   refreshAppMenu: () => { success: boolean };
   setDevMenuVisibility: (visible: boolean) => { success: boolean };
-  showContextMenu: (type: string, data?: any) => { success: boolean };
+  showContextMenu: (params: { data?: any; type: string }) => { success: boolean };
 }

package/packages/types/package.json

@@ -13,5 +13,8 @@
         "./src/*/index.ts"
       ]
     }
+  },
+  "dependencies": {
+    "@lobehub/chat-plugin-sdk": "^1.32.4"
   }
 }

package/packages/types/src/agent/chatConfig.ts

@@ -1,7 +1,7 @@
 /* eslint-disable sort-keys-fix/sort-keys-fix, typescript-sort-keys/interface */
 import { z } from 'zod';
 
-import { SearchMode } from '@/types/search';
+import { SearchMode } from '../search';
 
 export interface WorkingModel {
   model: string;

package/packages/types/src/agent/index.ts

@@ -1,7 +1,6 @@
-import { FileItem } from '@/types/files';
-import { KnowledgeBaseItem } from '@/types/knowledgeBase';
-import { FewShots, LLMParams } from '@/types/llm';
-
+import { FileItem } from '../files';
+import { KnowledgeBaseItem } from '../knowledgeBase';
+import { FewShots, LLMParams } from '../llm';
 import { LobeAgentChatConfig } from './chatConfig';
 
 export type TTSServer = 'openai' | 'edge' | 'microsoft';

package/packages/types/src/discover/assistants.ts

@@ -1,6 +1,6 @@
-import { FewShots } from '@/types/llm';
-import { MetaData } from '@/types/meta';
-import { LobeAgentSettings } from '@/types/session';
+import { FewShots } from '../llm';
+import { MetaData } from '../meta';
+import { LobeAgentSettings } from '../session';
 
 export enum AssistantCategory {
   Academic = 'academic',

package/packages/types/src/message/chat.ts

@@ -1,11 +1,11 @@
 import { IPluginErrorType } from '@lobehub/chat-plugin-sdk';
 
 import { ILobeAgentRuntimeErrorType } from '@/libs/model-runtime';
-import { ErrorType } from '@/types/fetch';
-import { MetaData } from '@/types/meta';
-import { MessageSemanticSearchChunk } from '@/types/rag';
-import { GroundingSearch } from '@/types/search';
 
+import { ErrorType } from '../fetch';
+import { MetaData } from '../meta';
+import { MessageSemanticSearchChunk } from '../rag';
+import { GroundingSearch } from '../search';
 import { MessageMetadata, MessageRoleType, ModelReasoning } from './base';
 import { ChatImageItem } from './image';
 import { ChatPluginPayload, ChatToolPayload } from './tools';

package/src/app/(backend)/_deprecated/createBizOpenAI/auth.ts

@@ -1,5 +1,6 @@
+import { ChatErrorType } from '@lobechat/types/fetch';
+
 import { getAppConfig } from '@/envs/app';
-import { ChatErrorType } from '@/types/fetch';
 
 interface AuthConfig {
   accessCode?: string | null;

package/src/app/(backend)/_deprecated/createBizOpenAI/createAzureOpenai.ts

@@ -1,8 +1,8 @@
+import { ChatErrorType } from '@lobechat/types/fetch';
 import OpenAI, { ClientOptions } from 'openai';
 import urlJoin from 'url-join';
 
 import { getLLMConfig } from '@/config/llm';
-import { ChatErrorType } from '@/types/fetch';
 
 // create Azure OpenAI Instance
 export const createAzureOpenai = (params: {

package/src/app/(backend)/_deprecated/createBizOpenAI/createOpenai.ts

@@ -1,7 +1,7 @@
+import { ChatErrorType } from '@lobechat/types/fetch';
 import OpenAI from 'openai';
 
 import { getLLMConfig } from '@/config/llm';
-import { ChatErrorType } from '@/types/fetch';
 
 // create OpenAI instance
 export const createOpenai = (userApiKey: string | null, endpoint?: string | null) => {

package/src/app/(backend)/_deprecated/createBizOpenAI/index.ts

@@ -1,7 +1,7 @@
+import { ChatErrorType, ErrorType } from '@lobechat/types/fetch';
 import OpenAI from 'openai';
 
 import { getOpenAIAuthFromRequest } from '@/const/fetch';
-import { ChatErrorType, ErrorType } from '@/types/fetch';
 import { createErrorResponse } from '@/utils/errorResponse';
 
 import { checkAuth } from './auth';

package/src/app/(backend)/middleware/auth/index.test.ts

@@ -1,7 +1,7 @@
+import { ChatErrorType } from '@lobechat/types/fetch';
 import { afterEach, beforeEach, describe, expect, it, vi } from 'vitest';
 
 import { AgentRuntimeError } from '@/libs/model-runtime';
-import { ChatErrorType } from '@/types/fetch';
 import { createErrorResponse } from '@/utils/errorResponse';
 import { getXorPayload } from '@/utils/server/xor';
 

package/src/app/(backend)/middleware/auth/index.ts

@@ -1,4 +1,5 @@
 import { AuthObject } from '@clerk/backend';
+import { ChatErrorType } from '@lobechat/types/fetch';
 import { NextRequest } from 'next/server';
 
 import {
@@ -11,7 +12,6 @@ import {
 import { ClerkAuth } from '@/libs/clerk-auth';
 import { AgentRuntimeError, ChatCompletionErrorPayload, ModelRuntime } from '@/libs/model-runtime';
 import { validateOIDCJWT } from '@/libs/oidc-provider/jwt';
-import { ChatErrorType } from '@/types/fetch';
 import { createErrorResponse } from '@/utils/errorResponse';
 import { getXorPayload } from '@/utils/server/xor';
 

package/src/app/(backend)/middleware/auth/utils.ts

@@ -1,9 +1,9 @@
 import { type AuthObject } from '@clerk/backend';
+import { ChatErrorType } from '@lobechat/types/fetch';
 
 import { enableClerk, enableNextAuth } from '@/const/auth';
 import { getAppConfig } from '@/envs/app';
 import { AgentRuntimeError } from '@/libs/model-runtime';
-import { ChatErrorType } from '@/types/fetch';
 
 interface CheckAuthParams {
   accessCode?: string;

package/src/app/(backend)/webapi/chat/[provider]/route.test.ts

@@ -1,11 +1,11 @@
 // @vitest-environment node
 import { getAuth } from '@clerk/nextjs/server';
+import { ChatErrorType } from '@lobechat/types/fetch';
 import { afterEach, beforeEach, describe, expect, it, vi } from 'vitest';
 
 import { checkAuthMethod } from '@/app/(backend)/middleware/auth/utils';
 import { LOBE_CHAT_AUTH_HEADER, OAUTH_AUTHORIZED } from '@/const/auth';
 import { LobeRuntimeAI, ModelRuntime } from '@/libs/model-runtime';
-import { ChatErrorType } from '@/types/fetch';
 import { getXorPayload } from '@/utils/server/xor';
 
 import { POST } from './route';

package/src/app/(backend)/webapi/chat/[provider]/route.ts

@@ -1,3 +1,5 @@
+import { ChatErrorType } from '@lobechat/types/fetch';
+
 import { checkAuth } from '@/app/(backend)/middleware/auth';
 import {
   AGENT_RUNTIME_ERROR_SET,
@@ -5,7 +7,6 @@ import {
   ModelRuntime,
 } from '@/libs/model-runtime';
 import { createTraceOptions, initModelRuntimeWithUserPayload } from '@/server/modules/ModelRuntime';
-import { ChatErrorType } from '@/types/fetch';
 import { ChatStreamPayload } from '@/types/openai/chat';
 import { createErrorResponse } from '@/utils/errorResponse';
 import { getTracePayload } from '@/utils/trace';

package/src/app/(backend)/webapi/models/[provider]/pull/route.ts

@@ -1,7 +1,8 @@
+import { ChatErrorType } from '@lobechat/types/fetch';
+
 import { checkAuth } from '@/app/(backend)/middleware/auth';
 import { ChatCompletionErrorPayload, PullModelParams } from '@/libs/model-runtime';
 import { initModelRuntimeWithUserPayload } from '@/server/modules/ModelRuntime';
-import { ChatErrorType } from '@/types/fetch';
 import { createErrorResponse } from '@/utils/errorResponse';
 
 export const runtime = 'edge';

package/src/app/(backend)/webapi/models/[provider]/route.ts

@@ -1,9 +1,9 @@
+import { ChatErrorType } from '@lobechat/types/fetch';
 import { NextResponse } from 'next/server';
 
 import { checkAuth } from '@/app/(backend)/middleware/auth';
 import { ChatCompletionErrorPayload, ModelProvider } from '@/libs/model-runtime';
 import { initModelRuntimeWithUserPayload } from '@/server/modules/ModelRuntime';
-import { ChatErrorType } from '@/types/fetch';
 import { createErrorResponse } from '@/utils/errorResponse';
 
 export const runtime = 'edge';

package/src/app/(backend)/webapi/plugin/gateway/route.ts

@@ -1,3 +1,4 @@
+import { ChatErrorType, ErrorType } from '@lobechat/types/fetch';
 import { PluginRequestPayload } from '@lobehub/chat-plugin-sdk';
 import { createGatewayOnEdgeRuntime } from '@lobehub/chat-plugins-gateway';
 
@@ -6,7 +7,6 @@ import { LOBE_CHAT_TRACE_ID, TraceNameMap } from '@/const/trace';
 import { getAppConfig } from '@/envs/app';
 import { AgentRuntimeError } from '@/libs/model-runtime';
 import { TraceClient } from '@/libs/traces';
-import { ChatErrorType, ErrorType } from '@/types/fetch';
 import { createErrorResponse } from '@/utils/errorResponse';
 import { getXorPayload } from '@/utils/server/xor';
 import { getTracePayload } from '@/utils/trace';

package/src/app/(backend)/webapi/text-to-image/[provider]/route.ts

@@ -1,10 +1,10 @@
+import { ChatErrorType } from '@lobechat/types/fetch';
 import { NextResponse } from 'next/server';
 
 import { checkAuth } from '@/app/(backend)/middleware/auth';
 import { ChatCompletionErrorPayload } from '@/libs/model-runtime';
 import { TextToImagePayload } from '@/libs/model-runtime/types';
 import { initModelRuntimeWithUserPayload } from '@/server/modules/ModelRuntime';
-import { ChatErrorType } from '@/types/fetch';
 import { createErrorResponse } from '@/utils/errorResponse';
 
 export const runtime = 'edge';

package/src/app/[variants]/(main)/settings/provider/features/ModelList/DisabledModels.tsx

@@ -1,7 +1,7 @@
 import { Button, Text } from '@lobehub/ui';
 import isEqual from 'fast-deep-equal';
 import { ChevronDown } from 'lucide-react';
-import { memo, useState } from 'react';
+import { memo, useMemo, useState } from 'react';
 import { useTranslation } from 'react-i18next';
 import { Flexbox } from 'react-layout-kit';
 
@@ -10,16 +10,26 @@ import { aiModelSelectors } from '@/store/aiInfra/selectors';
 
 import ModelItem from './ModelItem';
 
-const DisabledModels = memo(() => {
+interface DisabledModelsProps {
+  activeTab: string;
+}
+
+const DisabledModels = memo<DisabledModelsProps>(({ activeTab }) => {
   const { t } = useTranslation('modelProvider');
 
   const [showMore, setShowMore] = useState(false);
   const disabledModels = useAiInfraStore(aiModelSelectors.disabledAiProviderModelList, isEqual);
 
-  const displayModels = showMore ? disabledModels : disabledModels.slice(0, 10);
+  // Filter models based on active tab
+  const filteredDisabledModels = useMemo(() => {
+    if (activeTab === 'all') return disabledModels;
+    return disabledModels.filter((model) => model.type === activeTab);
+  }, [disabledModels, activeTab]);
+
+  const displayModels = showMore ? filteredDisabledModels : filteredDisabledModels.slice(0, 10);
 
   return (
-    disabledModels.length > 0 && (
+    filteredDisabledModels.length > 0 && (
       <Flexbox>
         <Text style={{ fontSize: 12, marginTop: 8 }} type={'secondary'}>
           {t('providerModels.list.disabled')}
@@ -27,7 +37,7 @@ const DisabledModels = memo(() => {
         {displayModels.map((item) => (
           <ModelItem {...item} key={item.id} />
         ))}
-        {!showMore && disabledModels.length > 10 && (
+        {!showMore && filteredDisabledModels.length > 10 && (
           <Button
             block
             icon={ChevronDown}

package/src/app/[variants]/(main)/settings/provider/features/ModelList/EnabledModelList/index.tsx

@@ -1,7 +1,7 @@
 import { ActionIcon, Text } from '@lobehub/ui';
 import isEqual from 'fast-deep-equal';
 import { ArrowDownUpIcon, ToggleLeft } from 'lucide-react';
-import { useState } from 'react';
+import { useMemo, useState } from 'react';
 import { useTranslation } from 'react-i18next';
 import { Center, Flexbox } from 'react-layout-kit';
 
@@ -11,7 +11,11 @@ import { aiModelSelectors } from '@/store/aiInfra/selectors';
 import ModelItem from '../ModelItem';
 import SortModelModal from '../SortModelModal';
 
-const EnabledModelList = () => {
+interface EnabledModelListProps {
+  activeTab: string;
+}
+
+const EnabledModelList = ({ activeTab }: EnabledModelListProps) => {
   const { t } = useTranslation('modelProvider');
 
   const enabledModels = useAiInfraStore(aiModelSelectors.enabledAiProviderModelList, isEqual);
@@ -20,6 +24,14 @@ const EnabledModelList = () => {
   const [batchLoading, setBatchLoading] = useState(false);
 
   const isEmpty = enabledModels.length === 0;
+
+  // Filter models based on active tab
+  const filteredModels = useMemo(() => {
+    if (activeTab === 'all') return enabledModels;
+    return enabledModels.filter((model) => model.type === activeTab);
+  }, [enabledModels, activeTab]);
+
+  const isCurrentTabEmpty = filteredModels.length === 0;
   return (
     <>
       <Flexbox horizontal justify={'space-between'}>
@@ -63,17 +75,23 @@ const EnabledModelList = () => {
           />
         )}
       </Flexbox>
+
       {isEmpty ? (
         <Center padding={12}>
           <Text style={{ fontSize: 12 }} type={'secondary'}>
             {t('providerModels.list.enabledEmpty')}
           </Text>
         </Center>
+      ) : isCurrentTabEmpty ? (
+        <Center padding={12}>
+          <Text style={{ fontSize: 12 }} type={'secondary'}>
+            {t('providerModels.list.noModelsInCategory')}
+          </Text>
+        </Center>
       ) : (
         <Flexbox gap={2}>
-          {enabledModels.map(({ displayName, id, ...res }) => {
+          {filteredModels.map(({ displayName, id, ...res }) => {
             const label = displayName || id;
-
             return <ModelItem displayName={label as string} id={id as string} key={id} {...res} />;
           })}
         </Flexbox>

package/src/app/[variants]/(main)/settings/provider/features/ModelList/index.tsx

@@ -1,7 +1,18 @@
 'use client';
 
+import { Icon, Tabs } from '@lobehub/ui';
 import { useTheme } from 'antd-style';
-import { Suspense, memo } from 'react';
+import isEqual from 'fast-deep-equal';
+import {
+  AudioLines,
+  BoltIcon,
+  Grid3x3Icon,
+  ImageIcon,
+  MessageSquareTextIcon,
+  MicIcon,
+} from 'lucide-react';
+import { Suspense, memo, useMemo, useState } from 'react';
+import { useTranslation } from 'react-i18next';
 import { Flexbox } from 'react-layout-kit';
 
 import { useIsMobile } from '@/hooks/useIsMobile';
@@ -20,14 +31,92 @@ interface ContentProps {
 }
 
 const Content = memo<ContentProps>(({ id }) => {
+  const { t } = useTranslation('modelProvider');
+  const [activeTab, setActiveTab] = useState('all');
+
   const [isSearching, isEmpty, useFetchAiProviderModels] = useAiInfraStore((s) => [
     !!s.modelSearchKeyword,
     aiModelSelectors.isEmptyAiProviderModelList(s),
     s.useFetchAiProviderModels,
   ]);
 
+  const allModels = useAiInfraStore(aiModelSelectors.filteredAiProviderModelList, isEqual);
+
   const { isLoading } = useFetchAiProviderModels(id);
 
+  // Count models by type (for all models, not just enabled)
+  const modelCounts = useMemo(() => {
+    const counts = {
+      all: allModels.length,
+      chat: 0,
+      embedding: 0,
+      image: 0,
+      stt: 0,
+      tts: 0,
+    };
+
+    allModels.forEach((model) => {
+      const type = model.type;
+      if (type && Object.prototype.hasOwnProperty.call(counts, type)) {
+        counts[type as keyof typeof counts]++;
+      }
+    });
+
+    return counts;
+  }, [allModels]);
+
+  // Tab definitions with counts (only show tabs with models > 0, except 'all' tab)
+  const tabs = useMemo(() => {
+    const formatTabLabel = (baseLabel: string, count: number) =>
+      count > 0 ? `${baseLabel} (${count})` : baseLabel;
+
+    const allTabs = [
+      {
+        count: modelCounts.all,
+        icon: <Icon icon={Grid3x3Icon} size={16} />,
+        key: 'all',
+        label: formatTabLabel(t('providerModels.tabs.all'), modelCounts.all),
+      },
+      {
+        count: modelCounts.chat,
+        icon: <Icon icon={MessageSquareTextIcon} size={16} />,
+        key: 'chat',
+        label: formatTabLabel(t('providerModels.tabs.chat'), modelCounts.chat),
+      },
+      {
+        count: modelCounts.image,
+        icon: <Icon icon={ImageIcon} size={16} />,
+        key: 'image',
+        label: formatTabLabel(t('providerModels.tabs.image'), modelCounts.image),
+      },
+      {
+        count: modelCounts.embedding,
+        icon: <Icon icon={BoltIcon} size={16} />,
+        key: 'embedding',
+        label: formatTabLabel(t('providerModels.tabs.embedding'), modelCounts.embedding),
+      },
+      {
+        count: modelCounts.stt,
+        icon: <Icon icon={MicIcon} size={16} />,
+        key: 'stt',
+        label: formatTabLabel(t('providerModels.tabs.stt'), modelCounts.stt),
+      },
+      {
+        count: modelCounts.tts,
+        icon: <Icon icon={AudioLines} size={16} />,
+        key: 'tts',
+        label: formatTabLabel(t('providerModels.tabs.tts'), modelCounts.tts),
+      },
+    ];
+
+    // Only show tabs that have models (count > 0), but always show 'all' tab
+    return allTabs.filter((tab) => tab.key === 'all' || tab.count > 0);
+  }, [modelCounts]);
+
+  // Ensure active tab is available, fallback to 'all' if current tab is hidden
+  const availableTabKeys = tabs.map((tab) => tab.key);
+  const currentActiveTab = availableTabKeys.includes(activeTab) ? activeTab : 'all';
+
   if (isLoading) return <SkeletonList />;
 
   if (isSearching) return <SearchResult />;
@@ -36,8 +125,15 @@ const Content = memo<ContentProps>(({ id }) => {
     <EmptyModels provider={id} />
   ) : (
     <Flexbox>
-      <EnabledModelList />
-      <DisabledModels />
+      <Tabs
+        activeKey={currentActiveTab}
+        items={tabs}
+        onChange={setActiveTab}
+        size="small"
+        style={{ marginBottom: 12 }}
+      />
+      <EnabledModelList activeTab={currentActiveTab} />
+      <DisabledModels activeTab={currentActiveTab} />
     </Flexbox>
   );
 });

package/src/config/modelProviders/ai302.ts

@@ -2,6 +2,7 @@ import { ModelProviderCard } from '@/types/llm';
 
 // ref: https://302.ai/pricing/
 const Ai302: ModelProviderCard = {
+  apiKeyUrl: 'https://lobe.li/Oizw5sN',
   chatModels: [
     {
       contextWindowTokens: 32_000,