Hi everyone! 👋

I'm from the Qwen Code team. I wanted to share some updates about the Qwen OAuth integration in OpenClaw and discuss the best path forward together.

A bit of background

First off, I want to say how amazing it's been to see OpenClaw grow and become so popular! The community's enthusiasm is incredible, and we're grateful that Qwen models have been part of that journey.

The Qwen OAuth integration was originally created by the community (through reverse-engineering our open-source code at https://github.com/QwenLM/qwen-code). While we didn't officially build or maintain this integration, we initially took a hands-off approach and just set a modest rate limit of 50 free calls per user per day.

The challenge we're facing

Here's the thing—OpenClaw has become so successful that we're running into some growing pains:

• When using Qwen OAuth through OpenClaw, users are frequently hitting HTTP 429 (rate limit) errors on qwen/coder-model, which isn't a great experience
• The traffic volume from OpenClaw has grown significantly, and it's starting to affect our ability to serve the broader Qwen Code user base, #2426

We want to be good stewards of both communities, so we need to make some changes.

What's happening

We're planning to restrict Qwen OAuth authentication from OpenClaw clients on the server side. The proposed timeline is:

The recommended path forward

We now officially support Alibaba Cloud Model Studio (Coding Plan), which is the recommended way to use Qwen models with OpenClaw:

1. Sign up at Alibaba Cloud Model Studio

2. Create a Coding Plan API key

3. Run:

   # China endpoint
   openclaw onboard --auth-choice modelstudio-api-key-cn
   
   # Global endpoint
   openclaw onboard --auth-choice modelstudio-api-key

Full docs: https://docs.openclaw.ai/providers/qwen

Why Model Studio?

• ✅ Officially supported by the Qwen team
• ✅ No rate limiting issues
• ✅ Access to Qwen 3.5 Plus, GLM-4.7, Kimi K2.5, MiniMax M2.5, and more
• ✅ Better reliability

Let's discuss

We genuinely want to hear your thoughts:

• Any concerns about the migration?
• Questions about setting up Model Studio?

Thanks for being such an amazing community! 🙏