Merge b9e3f47243 into 3809375694

Support OpenAI o3 and o4-mini

README.md

@@ -7,7 +7,7 @@
 
 
 
-<h1 align="center">NextChat (ChatGPT Next Web)</h1>
+<h1 align="center">NextChat</h1>
 
 English / [简体中文](./README_CN.md)
 
@@ -22,8 +22,7 @@ English / [简体中文](./README_CN.md)
 [![MacOS][MacOS-image]][download-url]
 [![Linux][Linux-image]][download-url]
 
-[NextChatAI](https://nextchat.dev/chat?utm_source=readme) / [Web App Demo](https://app.nextchat.dev) / [Desktop App](https://github.com/Yidadaa/ChatGPT-Next-Web/releases) 
-[NextChatAI](https://nextchat.club?utm_source=readme) / [Web App Demo](https://app.nextchat.dev) / [Desktop App](https://github.com/Yidadaa/ChatGPT-Next-Web/releases) / [Discord](https://discord.gg/YCkeafCafC) / [Enterprise Edition](#enterprise-edition) / [Twitter](https://twitter.com/NextChatDev)
+[NextChatAI](https://nextchat.club?utm_source=readme) / [iOS APP](https://apps.apple.com/us/app/nextchat-ai/id6743085599) / [Web App Demo](https://app.nextchat.dev) / [Desktop App](https://github.com/Yidadaa/ChatGPT-Next-Web/releases) / [Enterprise Edition](#enterprise-edition) 
 
 
 [saas-url]: https://nextchat.club?utm_source=readme
@@ -41,31 +40,14 @@ English / [简体中文](./README_CN.md)
 
 </div>
 
-## 👋 Hey, NextChat is going to develop a native app!
+## 🥳 Cheer for NextChat iOS Version Online!
+> [👉 Click Here to Install Now](https://apps.apple.com/us/app/nextchat-ai/id6743085599)
 
-> This week we are going to start working on iOS and Android APP, and we want to find some reliable friends to do it together!
+> [❤️ Source Code Coming Soon](https://github.com/ChatGPTNextWeb/NextChat-iOS)
 
+![Github iOS Image](https://github.com/user-attachments/assets/e0aa334f-4c13-4dc9-8310-e3b09fa4b9f3)
 
-✨ Several key points:
-
-- Starting from 0, you are a veteran
-- Completely open source, not hidden
-- Native development, pursuing the ultimate experience
-
-Will you come and do something together? 😎
-
-https://github.com/ChatGPTNextWeb/NextChat/issues/6269
-
-#Seeking for talents is thirsty #lack of people
-
-
-## 🥳 Cheer for DeepSeek, China's AI star!
- > Purpose-Built UI for DeepSeek Reasoner Model
  
-<img src="https://github.com/user-attachments/assets/f3952210-3af1-4dc0-9b81-40eaa4847d9a"/>
-
-
-
 ## 🫣 NextChat Support MCP  ! 
 > Before build, please set env ENABLE_MCP=true
 

app/client/api.ts

@@ -40,6 +40,11 @@ export interface MultimodalContent {
   };
 }
 
+export interface MultimodalContentForAlibaba {
+  text?: string;
+  image?: string;
+}
+
 export interface RequestMessage {
   role: MessageRole;
   content: string | MultimodalContent[];

app/client/platforms/alibaba.ts

@@ -7,7 +7,10 @@ import {
   ChatMessageTool,
   usePluginStore,
 } from "@/app/store";
-import { streamWithThink } from "@/app/utils/chat";
+import {
+  preProcessImageContentForAlibabaDashScope,
+  streamWithThink,
+} from "@/app/utils/chat";
 import {
   ChatOptions,
   getHeaders,
@@ -15,12 +18,14 @@ import {
   LLMModel,
   SpeechOptions,
   MultimodalContent,
+  MultimodalContentForAlibaba,
 } from "../api";
 import { getClientConfig } from "@/app/config/client";
 import {
   getMessageTextContent,
   getMessageTextContentWithoutThinking,
   getTimeoutMSByModel,
+  isVisionModel,
 } from "@/app/utils";
 import { fetch } from "@/app/utils/stream";
 
@@ -89,14 +94,6 @@ export class QwenApi implements LLMApi {
   }
 
   async chat(options: ChatOptions) {
-    const messages = options.messages.map((v) => ({
-      role: v.role,
-      content:
-        v.role === "assistant"
-          ? getMessageTextContentWithoutThinking(v)
-          : getMessageTextContent(v),
-    }));
-
     const modelConfig = {
       ...useAppConfig.getState().modelConfig,
       ...useChatStore.getState().currentSession().mask.modelConfig,
@@ -105,6 +102,21 @@ export class QwenApi implements LLMApi {
       },
     };
 
+    const visionModel = isVisionModel(options.config.model);
+
+    const messages: ChatOptions["messages"] = [];
+    for (const v of options.messages) {
+      const content = (
+        visionModel
+          ? await preProcessImageContentForAlibabaDashScope(v.content)
+          : v.role === "assistant"
+          ? getMessageTextContentWithoutThinking(v)
+          : getMessageTextContent(v)
+      ) as any;
+
+      messages.push({ role: v.role, content });
+    }
+
     const shouldStream = !!options.config.stream;
     const requestPayload: RequestPayload = {
       model: modelConfig.model,
@@ -129,7 +141,7 @@ export class QwenApi implements LLMApi {
         "X-DashScope-SSE": shouldStream ? "enable" : "disable",
       };
 
-      const chatPath = this.path(Alibaba.ChatPath);
+      const chatPath = this.path(Alibaba.ChatPath(modelConfig.model));
       const chatPayload = {
         method: "POST",
         body: JSON.stringify(requestPayload),
@@ -162,7 +174,7 @@ export class QwenApi implements LLMApi {
             const json = JSON.parse(text);
             const choices = json.output.choices as Array<{
               message: {
-                content: string | null;
+                content: string | null | MultimodalContentForAlibaba[];
                 tool_calls: ChatMessageTool[];
                 reasoning_content: string | null;
               };
@@ -212,7 +224,9 @@ export class QwenApi implements LLMApi {
             } else if (content && content.length > 0) {
               return {
                 isThinking: false,
-                content: content,
+                content: Array.isArray(content)
+                  ? content.map((item) => item.text).join(",")
+                  : content,
               };
             }
 

app/client/platforms/openai.ts

@@ -198,7 +198,8 @@ export class ChatGPTApi implements LLMApi {
     const isDalle3 = _isDalle3(options.config.model);
     const isO1OrO3 =
       options.config.model.startsWith("o1") ||
-      options.config.model.startsWith("o3");
+      options.config.model.startsWith("o3") ||
+      options.config.model.startsWith("o4-mini");
     if (isDalle3) {
       const prompt = getMessageTextContent(
         options.messages.slice(-1)?.pop() as any,
@@ -243,7 +244,7 @@ export class ChatGPTApi implements LLMApi {
       }
 
       // add max_tokens to vision model
-      if (visionModel) {
+      if (visionModel && !isO1OrO3) {
         requestPayload["max_tokens"] = Math.max(modelConfig.max_tokens, 4000);
       }
     }

app/constant.ts

@@ -221,7 +221,12 @@ export const ByteDance = {
 
 export const Alibaba = {
   ExampleEndpoint: ALIBABA_BASE_URL,
-  ChatPath: "v1/services/aigc/text-generation/generation",
+  ChatPath: (modelName: string) => {
+    if (modelName.includes("vl") || modelName.includes("omni")) {
+      return "v1/services/aigc/multimodal-generation/generation";
+    }
+    return `v1/services/aigc/text-generation/generation`;
+  },
 };
 
 export const Tencent = {
@@ -412,6 +417,14 @@ export const KnowledgeCutOffDate: Record<string, string> = {
   "gpt-4-turbo": "2023-12",
   "gpt-4-turbo-2024-04-09": "2023-12",
   "gpt-4-turbo-preview": "2023-12",
+  "gpt-4.1": "2024-06",
+  "gpt-4.1-2025-04-14": "2024-06",
+  "gpt-4.1-mini": "2024-06",
+  "gpt-4.1-mini-2025-04-14": "2024-06",
+  "gpt-4.1-nano": "2024-06",
+  "gpt-4.1-nano-2025-04-14": "2024-06",
+  "gpt-4.5-preview": "2023-10",
+  "gpt-4.5-preview-2025-02-27": "2023-10",
   "gpt-4o": "2023-10",
   "gpt-4o-2024-05-13": "2023-10",
   "gpt-4o-2024-08-06": "2023-10",
@@ -453,6 +466,7 @@ export const DEFAULT_TTS_VOICES = [
 export const VISION_MODEL_REGEXES = [
   /vision/,
   /gpt-4o/,
+  /gpt-4\.1/,
   /claude-3/,
   /gemini-1\.5/,
   /gemini-exp/,
@@ -464,6 +478,8 @@ export const VISION_MODEL_REGEXES = [
   /^dall-e-3$/, // Matches exactly "dall-e-3"
   /glm-4v/,
   /vl/i,
+  /o3/,
+  /o4-mini/,
 ];
 
 export const EXCLUDE_VISION_MODEL_REGEXES = [/claude-3-5-haiku-20241022/];
@@ -480,6 +496,14 @@ const openaiModels = [
   "gpt-4-32k-0613",
   "gpt-4-turbo",
   "gpt-4-turbo-preview",
+  "gpt-4.1",
+  "gpt-4.1-2025-04-14",
+  "gpt-4.1-mini",
+  "gpt-4.1-mini-2025-04-14",
+  "gpt-4.1-nano",
+  "gpt-4.1-nano-2025-04-14",
+  "gpt-4.5-preview",
+  "gpt-4.5-preview-2025-02-27",
   "gpt-4o",
   "gpt-4o-2024-05-13",
   "gpt-4o-2024-08-06",
@@ -494,6 +518,8 @@ const openaiModels = [
   "o1-mini",
   "o1-preview",
   "o3-mini",
+  "o3",
+  "o4-mini",
 ];
 
 const googleModels = [
@@ -570,6 +596,9 @@ const alibabaModes = [
   "qwen-max-0403",
   "qwen-max-0107",
   "qwen-max-longcontext",
+  "qwen-omni-turbo",
+  "qwen-vl-plus",
+  "qwen-vl-max",
 ];
 
 const tencentModels = [

app/utils/chat.ts

@@ -3,7 +3,7 @@ import {
   UPLOAD_URL,
   REQUEST_TIMEOUT_MS,
 } from "@/app/constant";
-import { RequestMessage } from "@/app/client/api";
+import { MultimodalContent, RequestMessage } from "@/app/client/api";
 import Locale from "@/app/locales";
 import {
   EventStreamContentType,
@@ -70,8 +70,9 @@ export function compressImage(file: Blob, maxSize: number): Promise<string> {
   });
 }
 
-export async function preProcessImageContent(
+export async function preProcessImageContentBase(
   content: RequestMessage["content"],
+  transformImageUrl: (url: string) => Promise<{ [key: string]: any }>,
 ) {
   if (typeof content === "string") {
     return content;
@@ -81,7 +82,7 @@ export async function preProcessImageContent(
     if (part?.type == "image_url" && part?.image_url?.url) {
       try {
         const url = await cacheImageToBase64Image(part?.image_url?.url);
-        result.push({ type: part.type, image_url: { url } });
+        result.push(await transformImageUrl(url));
       } catch (error) {
         console.error("Error processing image URL:", error);
       }
@@ -92,6 +93,23 @@ export async function preProcessImageContent(
   return result;
 }
 
+export async function preProcessImageContent(
+  content: RequestMessage["content"],
+) {
+  return preProcessImageContentBase(content, async (url) => ({
+    type: "image_url",
+    image_url: { url },
+  })) as Promise<MultimodalContent[] | string>;
+}
+
+export async function preProcessImageContentForAlibabaDashScope(
+  content: RequestMessage["content"],
+) {
+  return preProcessImageContentBase(content, async (url) => ({
+    image: url,
+  }));
+}
+
 const imageCaches: Record<string, string> = {};
 export function cacheImageToBase64Image(imageUrl: string) {
   if (imageUrl.includes(CACHE_URL_PREFIX)) {

jest.config.ts

@@ -15,6 +15,8 @@ const config: Config = {
   moduleNameMapper: {
     "^@/(.*)$": "<rootDir>/$1",
   },
+  extensionsToTreatAsEsm: [".ts", ".tsx"],
+  injectGlobals: true,
 };
 
 // createJestConfig is exported this way to ensure that next/jest can load the Next.js config which is async

jest.setup.ts

@@ -1,24 +1,22 @@
 // Learn more: https://github.com/testing-library/jest-dom
 import "@testing-library/jest-dom";
+import { jest } from "@jest/globals";
 
 global.fetch = jest.fn(() =>
   Promise.resolve({
     ok: true,
     status: 200,
-    json: () => Promise.resolve({}),
+    json: () => Promise.resolve([]),
     headers: new Headers(),
     redirected: false,
     statusText: "OK",
     type: "basic",
     url: "",
-    clone: function () {
-      return this;
-    },
     body: null,
     bodyUsed: false,
     arrayBuffer: () => Promise.resolve(new ArrayBuffer(0)),
     blob: () => Promise.resolve(new Blob()),
     formData: () => Promise.resolve(new FormData()),
     text: () => Promise.resolve(""),
-  }),
+  } as Response),
 );

package.json

@@ -17,8 +17,8 @@
     "prompts": "node ./scripts/fetch-prompts.mjs",
     "prepare": "husky install",
     "proxy-dev": "sh ./scripts/init-proxy.sh && proxychains -f ./scripts/proxychains.conf yarn dev",
-    "test": "jest --watch",
-    "test:ci": "jest --ci"
+    "test": "node --no-warnings --experimental-vm-modules $(yarn bin jest) --watch",
+    "test:ci": "node --no-warnings --experimental-vm-modules $(yarn bin jest) --ci"
   },
   "dependencies": {
     "@fortaine/fetch-event-source": "^3.0.6",

test/vision-model-checker.test.ts

@@ -1,3 +1,4 @@
+import { jest } from "@jest/globals";
 import { isVisionModel } from "../app/utils";
 
 describe("isVisionModel", () => {
@@ -50,7 +51,7 @@ describe("isVisionModel", () => {
 
   test("should identify models from VISION_MODELS env var", () => {
     process.env.VISION_MODELS = "custom-vision-model,another-vision-model";
-    
+
     expect(isVisionModel("custom-vision-model")).toBe(true);
     expect(isVisionModel("another-vision-model")).toBe(true);
     expect(isVisionModel("unrelated-model")).toBe(false);
@@ -64,4 +65,4 @@ describe("isVisionModel", () => {
     expect(isVisionModel("unrelated-model")).toBe(false);
     expect(isVisionModel("gpt-4-vision")).toBe(true);
   });
-});
+});