Code

Qwen3-Coder 480B A35B Instruct

480B-parameter MoE coding model with 35B active, 256K context, and agentic performance rivaling Claude Sonnet on complex development tasks.

Try our Qwen3-Coder API

read docs

About model

Qwen3-Coder-480B-A35B-Instruct excels at agentic coding tasks, achieving high performance on coding and browser-use challenges. It supports long-context capabilities and repository-scale understanding. Suitable for developers and power users.

Quickstart guides

Agents

How to Build Coding Agents

RAG

Building a RAG Workflow

Integrations

Using Together with Vercel's AI SDK

API usage

cURL
Python
Typescript

Endpoint:

Qwen/Qwen3-Coder-480B-A35B-Instruct-FP8

curl -X POST https://api.together.xyz/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $TOGETHER_API_KEY" \
  -d '{
    "model": "Qwen/Qwen3-Coder-480B-A35B-Instruct-FP8",
    "messages": [{
      "role": "user",
      "content": "Given two binary strings `a` and `b`, return their sum as a binary string"
    }]
  }'

from together import Together

client = Together()
response = client.chat.completions.create(
  model="Qwen/Qwen3-Coder-480B-A35B-Instruct-FP8",
  messages=[
  	{
	    "role": "user", 
      "content": "Given two binary strings `a` and `b`, return their sum as a binary string"
    }
 ],
)

print(response.choices[0].message.content)

import Together from "together-ai";

const together = new Together();

async function main() {
  const response = await together.chat.completions.create({
    model: "Qwen/Qwen3-Coder-480B-A35B-Instruct-FP8",
    messages: [{
      role: "user",
      content: "Given two binary strings `a` and `b`, return their sum as a binary string"
    }]
  });
  
  console.log(response.choices[0]?.message?.content);
}

main();

Model card
Architecture Overview:
• Mixture-of-Experts architecture with 480B total parameters and 35B active parameters
• Native 256K token context window, extendable to 1M tokens using Yarn interpolation
• Advanced grouped query attention mechanism for memory efficiency

Training Methodology:
• Trained on 7.5 trillion high-quality tokens with 70% code ratio across 358 programming languages
• Sophisticated post-training with supervised fine-tuning and reinforcement learning workflows
• Constitutional AI training for safety alignment and responsible code generation

Performance Characteristics:
• State-of-the-art SWE-bench Verified results at 69.6%, comparable to Claude Sonnet 4
• Exceptional agentic coding capabilities with 37.5 score on complex autonomous workflows
• Superior tool use and browser automation performance for development tasks
‍
Applications & use cases
Agentic Development Workflows:
• Autonomous software engineering tasks spanning multiple files and services
• Legacy system modernization with comprehensive analysis and migration planning
• End-to-end feature development across backend APIs, frontend components, & databases

Advanced Code Generation:
• Repository-scale refactoring and architectural improvements
• Complex debugging and root cause analysis across distributed systems
• Code completion and fill-in-the-middle for development environments

Enterprise Integration:
• Custom development workflows with fine-tuning capabilities
• Production deployment automation and CI/CD pipeline generation
• Code review assistance and security vulnerability identification

Developer Tooling:
• Integration with platforms like Qwen Code, CLINE, and VS Code extensions
• Batch processing for large codebase analysis and transformation
• Real-time coding assistance with context-aware suggestions
‍

Related models

Model specifications

Model data

Model provider
Qwen
Type
Code
Main use cases
Chat
Coding Agents
Function Calling
Features
Function Calling
JSON Mode
Deployment
Serverless
On-Demand Dedicated
Endpoint
Qwen/Qwen3-Coder-480B-A35B-Instruct-FP8
Parameters
480B
Activated parameters
35B
Context length
256K
Input price
$2.00 / 1M tokens
Output price
$2.00 / 1M tokens
Input modalities
Text
Output modalities
Text

Released
July 22, 2025
Last updated
July 24, 2025
Quantization level
FP8
External link
Provider docs
Category
Code

Run in Playground

Quickstart docs

Deploy model

Qwen3-Coder 480B A35B Instruct

About model

API usage

Model card

Applications & use cases