Qwen Code Open source

What it does

Qwen Code is an open-source AI coding agent from Alibaba's QwenLM team that lives in your terminal and also ships as IDE plugins, a desktop app, and chat integrations. It is designed to be "agentic out of the box," handling multi-step coding tasks with memory and delegated sub-work rather than needing every step spelled out.

What is inside

• Auto-Memory and Auto-Skills so the agent retains useful context and reusable procedures across sessions
• SubAgents and Agent Teams for splitting work across specialized agents
• Computer Use and an "Agent Arena" for evaluating agent behavior
• MCP support for connecting external tools and data sources
• IDE plugins for VS Code, JetBrains and Zed, plus desktop and chat-app integrations

Works with

Supports multiple model protocols out of the box, including OpenAI, Anthropic, Gemini and Qwen APIs, and offers official plugins for VS Code, JetBrains IDEs and Zed.

How to install or connect

Documented install options include a bash script (Linux/macOS), a PowerShell script (Windows), npm (Node.js 22+ required), and Homebrew (macOS/Linux).

Maintenance and safety

Apache-2.0 licensed and backed by Alibaba's Qwen research division, with roughly 28k stars, 3.1k forks and nearly 10,000 commits — solid signs of an official, well-resourced, actively maintained project. It publishes transparent evaluation results on SWE-bench-style benchmarks rather than only marketing claims, and follows conventional open governance (issues, PRs, releases).

Who should use it

Developers who want a serious, vendor-backed alternative to Claude Code or Codex CLI that isn't locked to one model provider, especially teams who want first-class JetBrains or Zed integration alongside a terminal agent, or who specifically want to route work through Qwen's own open models. It's also a reasonable pick for organizations that want an Apache-2.0, self-hostable option with a large corporate backer, rather than depending solely on a single closed-source vendor's terminal agent for day-to-day development work across a mixed IDE and terminal toolchain.