Here’s what you need to know. Want to know how to use Qwen 3.8? It’s Alibaba’s latest open-source model — strong at coding and agents, cheap to run, and easy to plug into an AI agent like Hermes.
Here’s how to access it, what it’s good for, and how to get the most out of it.
Last updated: July 2026.
Key takeaways
- Qwen 3.8 is a top open-source model — great for coding and agentic work, and cheap.
- Use it via a free/cheap API, or plug it into Hermes as the brain.
- It’s one of my favourite open-source picks, alongside GLM 5.2 and Kimi K3.
- Get the ready-made setup in the AI Profit Boardroom.
What Is Qwen 3.8?
Qwen 3.8 is the newest model in Alibaba’s open-source Qwen family. It’s built for real work — coding, building and automating — not just chat, and because it’s open source it’s cheap (often free) to run. Among Chinese open-source models, Qwen is one of the two I reach for most.
How to Use Qwen 3.8 (Step by Step)
- Pick how you’ll run it — a cheap/free API, or locally once support lands.
- Grab your API key from your provider (OpenRouter is an easy route).
- Point your tool at it — a chat app, a coding CLI, or an agent.
- Give it a real task — build a page, write code, automate something.
- Iterate — refine your prompt and let it work.
Using Qwen 3.8 With Hermes
The most powerful way to use Qwen 3.8 is as the brain of an AI agent. Plug it into Hermes — the free, open-source agent — and it can build, automate and run tasks on your files, cheaply. Full setup in my Hermes + Qwen 3.8 guide.
What Qwen 3.8 Is Good For
- Coding — writing and editing real projects
- Building websites, tools and automations
- Agentic tasks inside Hermes
- High-volume work where a frontier model is overkill
Qwen 3.8 vs Other Models
Qwen 3.8 is a superb cheap all-rounder. For top quality, a frontier model like Opus 4.8 wins; for frontier-grade open source, Kimi K3 is excellent; for fast-and-cheap, GLM 5.2. Qwen sits right in the value sweet spot. See my best model for Hermes agent guide.
Get the Ready-Made Setup
The whole setup — Qwen wired into an agent with memory and workflows — is in the AI Profit Boardroom.
New here? Start free with my AI course and community (plus 1,000+ AI agents), or grab a free strategy session.
How to Get Qwen 3.8 for Free
You can run Qwen 3.8 without a big bill. The easiest free route is a provider like OpenRouter, where open-source models are often available free or for pennies — grab the API key and plug it in. Once local support lands, you’ll also be able to run it on your own machine for free after setup.
For volume work, that low cost is the whole point: you get strong output without frontier-model prices, which is exactly why Qwen is such a good fit for agents that run all day.
Common Mistakes When Using Qwen 3.8
- Using it for tasks a free model can’t handle — match the model to the job
- Expecting frontier polish — it’s a value model, not Opus 4.8
- Giving it a vague brief — clear instructions get far better output
- Ignoring it as an agent brain — that’s where it shines
Qwen 3.8 for Everyday Work
Where Qwen 3.8 really earns its place is the high-volume, everyday tasks — drafting, coding, research, automation — that don’t need a frontier model but add up fast. Run those through Qwen and save your expensive model for the handful of jobs that genuinely need it. That split keeps quality high and costs low, which is the whole game with open-source models.
FAQ
How do I use Qwen 3.8?
Run it via a cheap or free API (OpenRouter is easy), point your tool or agent at it, and give it a real task. The most powerful use is as the brain of an agent like Hermes.
Is Qwen 3.8 free?
It’s open source and often available free via APIs, and cheap to run otherwise — far cheaper than a frontier model for the same work.
What is Qwen 3.8 good for?
Coding, building websites and tools, automating tasks, and agentic work. It’s a strong, cheap all-rounder.
Can I use Qwen 3.8 with Hermes?
Yes — plug it in as the brain and Hermes can build and automate on your files cheaply. It’s one of the best value setups going.
Qwen 3.8 vs GLM 5.2?
Both are excellent cheap open-source models. Qwen is great for building; GLM 5.2 is fast for high-volume work. Many people use both.
The Bottom Line
Use Qwen 3.8 for coding and agentic work at a fraction of frontier cost. The full setup is in the AI Profit Boardroom.