Qwen Code - ModRouter

Qwen Code supports modelProviders entries for OpenAI, Anthropic, and Gemini-style SDKs. Use the ModRouter endpoint that matches the auth type.

OpenAI-Compatible Setup

Set the ModRouter root URL and API key in the environment that launches Qwen Code:

export MODROUTER_BASE_URL="https://api.modrouter.online"
export MODROUTER_API_KEY="<your-modrouter-virtual-key>"

Add ModRouter under modelProviders.openai:

{
  "modelProviders": {
    "openai": [
      {
        "id": "openai/gpt-4o-mini",
        "name": "GPT via ModRouter",
        "baseUrl": "https://api.modrouter.online/openai/v1",
        "envKey": "MODROUTER_API_KEY"
      },
      {
        "id": "anthropic/claude-sonnet-4-5",
        "name": "Claude via ModRouter OpenAI format",
        "baseUrl": "https://api.modrouter.online/openai/v1",
        "envKey": "MODROUTER_API_KEY"
      }
    ]
  }
}

Anthropic or Gemini Auth Types

If you configure Qwen Code’s anthropic or gemini auth type, use the matching ModRouter endpoint:

Qwen Code auth type	ModRouter base URL
`openai`	`https://api.modrouter.online/openai/v1`
`anthropic`	`https://api.modrouter.online/anthropic`
`gemini`	`https://api.modrouter.online/genai`

Model Selection

Use Qwen Code’s /model picker to select the configured ModRouter model entry. Prefer explicit provider/model IDs so ModRouter routing is unambiguous.