Qwen3-Coder¶

In this article

• Main features of Qwen3-Coder
• Deployment Features
• Getting started after deploying Qwen3-Coder
• Order a server with Qwen3-Coder via API

Main features of Qwen3-Coder¶

• Efficient 30B model: The qwen3-coder:30b model offers 30B total parameters with only 3.3B activated, providing high performance while maintaining efficiency;
• Exceptional agent capabilities: Optimized for real software development tasks through advanced reinforcement learning on long-term tasks using SWE-Bench and similar benchmarks;
• Long context support: Native support for 256K tokens with the ability to extend up to 1M tokens using scale-optimized extrapolation methods, optimized for understanding repository scales;
• Scaled pretraining: Trained on 7.5 trillion tokens with a 70% code ratio while maintaining strong general and mathematical capabilities;
• Execution-based learning: Reinforcement learning based on code execution significantly increases the success rate of executing code in various real programming tasks;
• Integration with Open Web UI: Provides a modern web interface for convenient interaction with the model through port 8080, ensuring full control over data and request processing;
• Security and control: Complete local deployment ensures code and data confidentiality, while OLLAMA_HOST and OLLAMA_ORIGINS settings guarantee network security;
• Fault tolerance: An integrated system automatically restarts containers and services to ensure stable operation.
• Usage examples:
  • Software development: Automating code writing, refactoring, and debugging;
  • Agent tasks: Executing complex multi-step programming tasks using tools;
  • Browser work: Automating web development and testing;
  • Repository analysis: Understanding and working with large codebases;
  • Code review: Automatically analyzing and improving code quality;
  • Code documentation: Generating technical documentation and comments.

Deployment Features¶

Technical characteristics of the build:

• Ubuntu 22.04 with kernel update to version 6;
• Latest NVIDIA drivers;
• CUDA Toolkit;
• Ollama for model management;

• OpenWebUI for web interface. Installation features:

• Installation time is 25-45 minutes including OS installation;

• The Ollama server loads and runs the Qwen3-Coder model in GPU/RAM memory;
• Open WebUI is deployed as a web application connected to the Ollama server;
• Users interact with the model through the Open WebUI web interface for programming and agent tasks;
• All computations and code processing occur locally on the server;
• Administrators can configure the model for specific development tasks through OpenWebUI tools;
• Support for various levels of quantization to optimize memory usage.

Getting started after deploying Qwen3-Coder¶

After payment, a notification about the readiness of the server for work will be sent to the email specified at registration. It will contain the VPS IP address, as well as login and password for connecting to the server and a link to access the OpenWebUI panel. Our company's clients manage equipment in the server management and API panel — Invapi.

• Authentication data for accessing the server's operating system (e.g., via SSH) will be sent to you in the e-mail.

• Link for accessing the Ollama management panel with Open WebUI web interface: in the webpanel tag in the Info >> Tags tab of the Invapi control panel. The exact link in the form of https://qwen3-coder<Server_ID_from_Invapi>.hostkey.in will be sent in the letter sent upon server deployment.

After clicking the link from the tag webpanel, a Get started with Open WebUI login window will open, where you need to create an admin account name, email, and password for your chatbot, then press the Create Admin Account button:

After successful registration, the main interface of Open WebUI with access to Qwen3-Coder will open:

Order a server with Qwen3-Coder via API¶

To install this software using the API, follow these instructions.

Some of the content on this page was created or translated using AI.