Run Qwen Code in MXterm: install, first session and a split pane for tests
To run Qwen Code in MXterm, install it once from Qwen's instructions, open an MXterm pane in your project, run qwen and give it a task. Keep a plain shell in a split pane next to it for your tests. MXterm does not install the agent or sign you in; Qwen's docs say the first interactive session prompts you to configure authentication.
The short version
1. Install it from Qwen's instructions.
2. Open your project folder in MXterm and open a terminal pane.
3. Run qwen in the pane, then split the pane for a shell.
Install it
Qwen's repository describes it as "The open-source AI coding agent for your terminal, editor, desktop, browser, and chat." (github.com/QwenLM/qwen-code, read 31 October 2026.) This page is about the terminal tool, not the Qwen 2.5 Coder model family.
The quickstart lists a quick-install script for Linux and macOS, marked as recommended, and a manual route: Node.js 22 or later with npm install -g @qwen-code/qwen-code@latest, or Homebrew with brew install qwen-code. You start it with qwen (qwenlm.github.io/qwen-code-docs/en/users/quickstart/, read 31 October 2026).
Commands change, so check the page on the day you install. MXterm does not install agents.
Step 1: open a pane and run qwen
Open MXterm on your project folder and open a terminal pane; it is a real shell. Run qwen there.
Because the pane is a real shell, the agent runs in it as it would in any terminal.
Step 2: choose how it connects, then give it a task
The quickstart says: "When you start an interactive session with the qwen command, you’ll be prompted to configure authentication". You can run /auth anytime to change the method. The first-run menu offers QwenCloud, Alibaba ModelStudio, Third-party Providers and Custom Provider, and describes the last as a way to "manually connect a local server, proxy, or unsupported provider" (same page, read 31 October 2026).
The same page notes: "Qwen OAuth was discontinued on April 15, 2026." Follow the current menu. Then type your task in the agent.
Step 3: a shell beside it
Use the split buttons in the pane header (Split Right and Split Down) to put a plain shell next to it for your tests. Split panes run real shells.
Run your test command in the shell pane, and keep the agent in the pane beside it.
Next to other agents
It can run in a pane beside Claude Code or Codex in the same workspace, and each agent can work on its own branch in its own worktree. That reduces collisions; it does not stop two agents editing the same file.
See Claude Code in MXterm, the guide to running Claude Code and Codex side by side in split panes and isolating agents with git worktrees. To keep agents in one room, see the Control Room. Our list of AI coding agents in 2026 gives the vendors' own descriptions of each one.
Related guideRun Claude Code and Codex in Split Terminal PanesWhen a plain terminal is enough
For a task with this agent, a plain terminal is enough once it is installed and connected.
MXterm helps when agents or tasks run in parallel and you want them in one window. To try it, use the download page.
Multi-pane SSH, Command Recipes (Cmd+J), and a multi-agent Control Room (formerly War Room): free to start.
Already have MXterm Pro? Read the setup guide →
New to MXterm Pro? See pricing →
Frequently Asked Questions
Its repository calls it "The open-source AI coding agent for your terminal, editor, desktop, browser, and chat." It is a coding agent you run, not the Qwen 2.5 Coder model family (github.com/QwenLM/qwen-code, read 31 October 2026).
The agent itself runs in your terminal. Where the model runs depends on what you connect: the quickstart describes a Custom Provider option to "manually connect a local server, proxy, or unsupported provider". This page does not claim a local-model setup (qwenlm.github.io/qwen-code-docs/en/users/quickstart/, read 31 October 2026).
Yes. Split the pane and keep a plain shell next to the agent for your tests. Split panes run real shells.
No. MXterm runs the CLIs you install and does not include a model or sign you in. The agent asks you to configure authentication the first time you run it.
Yes, in separate panes of the same workspace, for example beside Claude Code or Codex. Each agent can work on its own branch in its own worktree.