For the first time in AI history, the terminal has been liberated. The latest update to QwenCode allows developers to manage entire server-side repositories from their pocket. Whether you are on a train or at a café, you can now execute complex deployments using nothing but a messaging app.
1. Remote Control: Coding via Telegram & WeChat
The biggest “FlowHub” favorite in this update is the native IM Channel Integration. You no longer need a complex SSH setup or a laptop to interact with your code.
- Direct Interaction: By running
qwen channel configure-telegram(or WeChat/DingTalk), you link your local environment to a bot. - Seamless Workflow: You can send a message like “Fix the CSS bug on the landing page and push to main,” and the AI will execute the task on your server and report back with a success message.
- No Desktop Required: This turns your phone into a high-powered development workstation.
2. Qwen3.6-Plus: The 1-Million Token Giant
The core engine has been upgraded to Qwen3.6-Plus, which now features a massive 1,000,000 token context window.
- Repository-Scale Reasoning: You can now feed your entire codebase into a single prompt. The AI doesn’t just look at one file; it understands the cross-file dependencies of your entire project.
- Multimodal Perception: The model can now “see” your UI. If you send a screenshot of a broken layout via Telegram, Qwen3.6-Plus analyzes the image and the code simultaneously to provide a fix.
3. The “/plan” Mode: Architect First, Code Second
To prevent expensive mistakes in large-scale projects, Qwen introduced the Plan-Execute Loop:
/planCommand: Before writing a single line of code, the AI generates a step-by-step architectural plan.- Safety & Accuracy: It lists every file it will touch and every command it will run. You review the plan on your phone, hit “Approve,” and only then does the AI begin execution.
- Cron Job Scheduling: You can now schedule tasks directly in natural language (e.g., “Run the database cleanup script every Sunday at 2 AM”), and the system handles the crontab configuration automatically.
4. Dynamic Model Routing (Cost Optimization)
Qwen now uses Hybrid Routing to save you money. It analyzes your request and intelligently splits the work:
- Easy Tasks: Routed to the lightning-fast Qwen-Flash.
- Complex Architecture: Routed to the heavyweight Qwen3.6-Plus. This ensures you get the highest quality for complex logic without paying “Plus” prices for simple text edits.
More info about the release : https://github.com/QwenLM/qwen-code/releases