# Qwen Code

Connect → Qwen Code

# Use AxForge with Qwen Code

    Qwen Code is the Qwen team's
    coding agent for the terminal. It takes any OpenAI-compatible service from three environment variables, and it
    is at home on Qwen3.8 27B.

    **Before you start**

        - An AxForge API key. Get one (a minute).

        - Node.js 20 or newer. This guide was tried with Qwen Code 0.25.2 on 2026-10-17.

## Set it up

### Install Qwen Code

```
npm install -g @qwen-code/qwen-code
```

### Point it at AxForge

```
export OPENAI_API_KEY="orx_live_your_key_here"
export OPENAI_BASE_URL=https://api.axforge.ai/v1
export OPENAI_MODEL=qwen3.8-27b-nvfp4
```

```
$env:OPENAI_API_KEY = "orx_live_your_key_here"
$env:OPENAI_BASE_URL = "https://api.axforge.ai/v1"
$env:OPENAI_MODEL = "qwen3.8-27b-nvfp4"
```

They last until you close this terminal. Start Qwen Code from the same one.

### Tell it the model's size

Qwen Code takes our model for a bigger Qwen and would count on a million tokens. Ours read 262,144 at once.
        Create `~/.qwen/settings.json` (on Windows `%USERPROFILE%\.qwen\settings.json`):

```
{
  "model": {
    "generationConfig": { "contextWindowSize": 262144 }
  }
}
```

With it, Qwen Code shortens a long session before it reaches our limit.

### Start it in your project

```
cd my-project
qwen
```

The box at the top says **API Key | qwen3.8-27b-nvfp4**: it runs on your AxForge key.

### Give it a task

Ask in plain words. It reads the file, edits it and shows the result. In its default **Auto mode** it
        approves safe steps itself and stops risky ones; Shift+Tab changes that.

## The settings

        | Variable | Value |  |

        | OPENAI_API_KEY | your AxForge key |  |

        | OPENAI_BASE_URL | https://api.axforge.ai/v1 |  |

        | OPENAI_MODEL | qwen3.8-27b-nvfp4, glm-5.3-flash or mimo-v2.6-flash |  |

        | contextWindowSize (~/.qwen/settings.json) | 262144 |  |

Which model for what: `qwen3.8-27b-nvfp4` is the
    everyday default and reads images, `glm-5.3-flash` thinks before it answers,
    `mimo-v2.6-flash` is the fastest.

## Check that it works

    The box at the top says **API Key** and your model, the footer says **262.1k Context**, and a
    task gets done. The calls show in the
    console under API Keys → Recent calls within a minute.

## When it does not work

      | What you see | What it means | Fix |  |

        | It asks how to sign in, or "No auth type is selected" | `OPENAI_API_KEY` is not set in this terminal | set the three variables as in step 2, in the same terminal |  |

        | "API Error: 401 Invalid API key. Create one at https://console.axforge.ai/keys." | the key was refused | paste it again, whole, or make a new one |  |

    Test a call says in one click whether a problem is your key,
    your balance or us.

      ← LibreChat
      OpenAI SDK and curl →



Source: https://dev.axforge.ai/docs/connect/qwen-code/