262К

Out

Chat

disable

Qwen Max 2025-09-23

Qwen Max is a large MoE LLM excelling in chat, coding, and complex problem-solving, pre-trained on 20T tokens

Text to Speech

Javascript

Python

                                        const { OpenAI } = require('openai');

const api = new OpenAI({
  baseURL: 'https://api.ai.cc/v1',
  apiKey: '',
});

const main = async () => {
  const result = await api.chat.completions.create({
    model: 'qwen-max-2025-09-23',
    messages: [
      {
        role: 'system',
        content: 'You are an AI assistant who knows everything.',
      },
      {
        role: 'user',
        content: 'Tell me, why is the sky blue?'
      }
    ],
  });

  const message = result.choices[0].message.content;
  console.log(`Assistant: ${message}`);
};

main();

                                        import os
from openai import OpenAI

client = OpenAI(
    base_url="https://api.ai.cc/v1",
    api_key="",    
)

response = client.chat.completions.create(
    model="qwen-max-2025-09-23",
    messages=[
        {
            "role": "system",
            "content": "You are an AI assistant who knows everything.",
        },
        {
            "role": "user",
            "content": "Tell me, why is the sky blue?"
        },
    ],
)

message = response.choices[0].message.content

print(f"Assistant: {message}")

Docs

300+ AI Models for OpenClaw & AI Agents

Get API Key Explore Models

Qwen Max 2025-09-23

Product Detail

Discover Qwen Max 2025-09-23: Alibaba's Advanced MoE Language Model

Developed by Alibaba, Qwen Max 2025-09-23 stands as a cutting-edge, large-scale Mixture-of-Experts (MoE) language model designed for unparalleled performance in diverse AI applications. This powerful model excels in chat interactions, complex coding challenges, and intricate problem-solving tasks, setting a new standard for intelligent systems.

Its innovative Transformer-based MoE architecture incorporates expert routing, intelligently activating only a subset of parameters per token. This design ensures an optimal balance between high performance and impressive computational efficiency. With pre-training on over 20 trillion tokens from vast multilingual and domain-specific corpora, Qwen Max is primed for deep reasoning and advanced text generation.

Technical Specifications & Core Architecture

★ Architecture: Transformer-based Mixture-of-Experts (MoE) with selective expert activation for each input token.
★ Total Parameters: Tailored for large-scale MoE designs, focusing on efficient activation rather than total count.
★ Pre-training Data: Exceeds 20 trillion tokens, encompassing extensive English, Chinese, and various major world languages.
★ Context Window: An impressive 262K tokens natively, supporting long-form understanding and generation.
★ Language Support: Primarily optimized for English and Chinese, offering robust bilingual capabilities.
★ Training Methods: Utilizes Supervised Fine-Tuning (SFT) and Reinforcement Learning from Human Feedback (RLHF) for superior output quality.

Key Capabilities & Benefits

✔ Efficient MoE Design: Delivers strong performance with significantly reduced computational overhead, making it cost-effective.
✔ Bilingual Proficiency: Seamlessly supports both English and Chinese, opening broad application scenarios across global markets.
✔ Expert in Coding & Problem-Solving: Excels in generating code, tackling complex problems, and enhancing conversational AI experiences through sophisticated reasoning.
✔ Advanced Fine-Tuning: Employs sophisticated fine-tuning strategies to ensure high-quality generation and strong alignment with human preferences and instructions.

Performance Metrics

How to Access and Use Qwen Max

Qwen Max 2025-09-23 is readily available on the AI/ML API platform. Developers can integrate this powerful model into their applications using its designated name: "Qwen Max 2025-09-23".

💡 Access Qwen Max API

Get started with Qwen Max 2025-09-23 API by signing up on the platform.

📄 Detailed API Documentation

For comprehensive integration guidelines and code examples, refer to the official API Documentation.

Ethical Guidelines and Licensing

▶ Transparency: Alibaba is committed to transparency regarding Qwen Max's capabilities and inherent limitations.
▶ Responsible Use: Users are strongly encouraged to adhere to responsible and ethical practices to prevent misuse or the generation of harmful content.
▶ Licensing: Usage is governed by specific licensing terms provided by Alibaba Cloud. Users must review these permissions and restrictions thoroughly before deployment.

Frequently Asked Questions (FAQ)

Q1: What makes Qwen Max 2025-09-23 unique?

A: Its Transformer-based Mixture-of-Experts (MoE) architecture with expert routing allows for high performance by activating only a subset of parameters, ensuring both efficiency and powerful capabilities in chat, coding, and problem-solving.

Q2: What languages does Qwen Max primarily support?

A: Qwen Max is primarily optimized for both English and Chinese, offering robust bilingual proficiency for diverse application scenarios.

Q3: How extensive is its context window?

A: It boasts an impressive native context window of 262K tokens, enabling it to process and understand very long inputs and generate coherent, extended responses.

Q4: Where can I find the API for Qwen Max 2025-09-23?

A: The model is available on the AI/ML API platform. You can sign up there to get access to the API and integrate it into your projects.

Q5: Are there any ethical guidelines for using Qwen Max?

A: Yes, Alibaba emphasizes transparency and encourages responsible, ethical use to prevent misuse. Users should review the specific licensing terms provided by Alibaba Cloud for usage permissions and restrictions.

AI Playground

Test all API models in the sandbox environment before you integrate. We provide more than 300 models to integrate into your app.

300+ AI Models for
OpenClaw & AI Agents

Save 20% on Costs

Free $1 Tokens for New Members