Enterprise Plans
AI Application Development
AI Translator API
AI SEO/GEO Service
Web Scraping Service
Log in
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/Qwen2-72B-Instruct',
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/Qwen2-72B-Instruct",
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}")
-
AI Playground
Test all API models in the sandbox environment before you integrate.
We provide more than 300 models to integrate into your app.
.webp)
Product Detail
✨ Introducing Qwen 2 Instruct (72B): A New Era in Language Models
The highly anticipated Qwen 2 series has arrived, marking a significant leap forward from its predecessor, Qwen 1.5. The 72B Instruct variant stands out with enhanced capabilities across multiple dimensions, setting new benchmarks in linguistic proficiency and contextual understanding. This powerful model is engineered to deliver state-of-the-art results in a wide array of applications, making it an indispensable tool for developers and researchers alike.
🚀 Key Architectural & Performance Upgrades
Qwen 2 builds upon the robust Transformer architecture, integrating advanced features that optimize efficiency and focus:
- Expanded Linguistic Proficiency: The model now supports an impressive 27 additional languages, significantly broadening its global applicability.
- Massive Context Length: Qwen 2 boasts an impressive 128K tokens context length support, enabling more comprehensive and contextually rich interactions without "blind spots" on large inputs.
- Advanced Architecture: Incorporates innovations like SwiGLU activation, attention QKV bias, Group Query Attention (GQA), and a mixture of sliding window attention for superior performance.
- Universal Group Query Attention: A significant enhancement over Qwen 1.5, GQA is universally applied for faster speeds and reduced memory usage during inference.
- Improved Domain Expertise: Demonstrates enhanced performance in critical areas such as coding, mathematics, and complex reasoning tasks.
💡 Versatile Use Cases for Qwen 2 72B Instruct
The powerful capabilities of Qwen 2 72B Instruct make it suitable for a diverse range of applications:
Text Generation & Chatbots
Its substantial size and superior language generation abilities make Qwen 2 72B Instruct an excellent choice for developing highly engaging and contextually aware chatbots. It facilitates natural, human-like conversational experiences across various topics.
Retrieval Augmented Generation (RAG) & Function Calling
Qwen 2 excels in tasks requiring augmentation with retrieval-generation, providing reliable outputs with reduced hallucinations. Furthermore, its proficiency in utilizing external tools (like function calls) positions it as an ideal foundation for building sophisticated AI agents that interact effectively with the real world.
Content Moderation 🛡️
With powerful language understanding, Qwen 2 72B Instruct is highly effective for content moderation. It can identify and filter inappropriate content across multiple languages, competing favorably with top-tier models like GPT-4 in safety tasks and outperforming others like Mistral-8x22B.
Multilingual Applications 🌐
Demonstrating excellence across 12 languages, Qwen 2 72B Instruct has been extensively trained in over 27 languages (including Western European, Eastern European, Middle Eastern, Eastern Asian, South-Eastern Asian, Southern Asian). It adeptly handles complex contexts (up to 128K tokens via YARN) and effectively manages code-switching, proving invaluable for diverse linguistic and cultural challenges. As highlighted in our comparison, "Qwen 2 vs LLama 3 Comparison", it performs remarkably well even with subtle, culturally nuanced prompts, making it a must-have for working with Asian language groups.
📈 How Qwen 2 72B Instruct Stacks Up Against Competitors
Qwen 2 72B Instruct consistently outperforms competitors like Llama 3 70B on classic benchmarks such as GSM8K and MATH. In critical areas like fraud and pornography detection, its performance is comparable to industry leaders like GPT-4 and Mistral. Notably, Qwen 2 shows increased alignment with human preference on the AlignBench benchmark and offers a substantial 128K context window, positioning it as a top contender in the AI landscape.
⚙️ Getting Started with Qwen 2 72B Instruct
API Access: Integrate Qwen 2 into your applications by signing up for AI/ML API access on this website.
Local Deployment: To experiment with Qwen 2 72B Instruct locally, ensure you have the newest Huggingface Transformers library (version 4.37.0 or later). Pay attention to the generation_config.json file in the model's repository for configurations, especially concerning code-switching.
📄 License Agreement
Qwen 2-72B Instruct operates under the Tongyi Qianwen license. Full details can be found on its GitHub or Hugging Face repository. While commercial use is generally free, products or services with over 100 million monthly active users are required to contact the developers for explicit permission.
✅ Conclusion: The Power of Qwen 2
Qwen 2 decisively surpasses its predecessor, offering unparalleled multilingual capabilities across 27 languages and robust 128K context support for complex tasks. Its excellence spans diverse applications, from advanced chatbot development and critical content moderation to sophisticated multilingual processing. Benchmarking confirms its strong competitive standing, making Qwen 2 an invaluable asset for a wide array of AI requirements. Enhanced accessibility through both APIs and local installation options further solidifies its position as a leading-edge language model.
🖥️ API Example
📊 AI Model Comparisons
If you want to see how this model fares against other AIs check out this Comparison:
❓ Frequently Asked Questions (FAQ) about Qwen 2 72B Instruct
Q: What are the main improvements of Qwen 2 over Qwen 1.5?
A: Qwen 2 significantly expands linguistic support to 27 additional languages, boosts context length to an impressive 128K tokens, and incorporates advanced architectural features like universal Group Query Attention for enhanced efficiency and performance.
Q: Can Qwen 2 72B Instruct handle multilingual tasks effectively?
A: Absolutely. Qwen 2 is trained in over 27 languages, excels in multilingual contexts, and is particularly strong in handling code-switching and nuances in Asian language groups, making it highly versatile for global applications.
Q: What kind of applications is Qwen 2 72B Instruct best suited for?
A: Qwen 2 72B Instruct is ideal for sophisticated chatbots, Retrieval Augmented Generation (RAG), advanced function calling for AI agents, robust content moderation across languages, and a wide range of complex multilingual applications.
Q: How does Qwen 2 compare to other leading models like Llama 3 or GPT-4?
A: Qwen 2 72B Instruct generally outperforms Llama 3 70B on classic benchmarks and shows comparable performance to GPT-4 and Mistral in safety-critical tasks like fraud detection, while also offering a large 128K context window and superior human preference alignment.
Q: What are the licensing terms for commercial use of Qwen 2-72B Instruct?
A: Qwen 2-72B Instruct uses the Tongyi Qianwen license. It is free for commercial use, but if your product or service reaches over 100 million monthly active users, you are required to contact the developers for specific permission.
Learn how you can transformyour company with AICC APIs
