How to Use Qwen3.6 27B in Devin
Qwen3.6 27B is useful for code fixes, web apps, and STEM problems. Compared with Qwen3.5 27B, it improves agentic coding and STEM reasoning.
Prompts to try with Qwen3.6 27B in Devin
Copy an example into Devin, or ask in your own words.
Use Qwen3.6 27B to investigate the failing tests in this repository. Reproduce one failure, make a focused fix, and report the verification result.
Use Qwen3.6 27B to add keyboard navigation to the project's search results. Follow existing component conventions and verify focus behavior.
Use Qwen3.6 27B to trace checkout validation through this codebase. Identify where invalid quantities are accepted and propose a regression test.
Use Qwen3.6 27B to build a chart from the project's sample dataset. Label units, handle missing values, and check the displayed totals.
Use Qwen3.6 27B with temperature 0.6 to refactor the date parser. Preserve documented behavior and check leap days, invalid dates, and timezone boundaries.
How to set up Qwen3.6 27B in Devin
Pick the way that suits you. Each takes about two minutes.
Use Spiral's MCP server
- Add a custom MCP
Open Customize › MCPs. Click + and choose New MCP. Fill in:
- Sign in to Spiral and approve
In Customize › MCPs, open Spiral and choose Connect. Spiral opens: sign in, pick the organization and click Authorize. No API key.
- Check it
On Spiral's configuration page, click Test tools. Devin lists Spiral's tools.
Use the Spiral skill
- Send Devin this message
Your agent reads Spiral's skill and sets itself up.
set up https://spiral.ai/SKILL.md
- Sign in or paste a key
Approve Spiral in the browser, or give the agent a key from the console.
Use the Spiral CLI
- Install and sign in
Installs the Spiral CLI, signs you in and adds its skill to your coding agents.
npx -y spiral-cli@latest init
Call the Spiral API
- Create an API key
In the console, under API keys.
- Run Qwen3.6 27B
One request; the result comes back by polling or a webhook.
curl https://api.spiral.ai/v1/runs \ -H "Authorization: Bearer $SPIRAL_API_KEY" \ -H "Content-Type: application/json" \ -H "Idempotency-Key: $(uuidgen)" \ -d '{ "provider": "openrouter", "model": "qwen/qwen3.6-27b", "input": { "messages": [ { "role": "system" } ] } }'
Then paste one of the prompts above: Devin finds Qwen3.6 27B on Spiral, shows the price and runs it.
Qwen3.6 27B providers and prices
Spiral runs Qwen3.6 27B through OpenRouter. Open a card for today's price, with every fee included.
Up to $0.017055 per run at the default inputs. $0.416 per 1M input tokens, $4.16 per 1M output tokens.
Settings you can ask for
Ask Devin for any of these in your prompt.
| Setting | Allowed values | Default | Description |
|---|---|---|---|
| Seed | a whole number | ||
| Min p | 0 to 1 | ||
| Top k | a whole number from 0 | ||
| Top p | 0 to 1 | ||
| Logprobs | on or off | ||
| Max tokens | 1 to 65536 | The most output tokens to generate. | |
| Temperature | 0 to 2 | ||
| Top logprobs | 0 to 20 | ||
| Presence penalty | -2 to 2 | ||
| Frequency penalty | -2 to 2 | ||
| Repetition penalty | 0 to 2 |
Every API field of Qwen: Qwen3.6 27B is on its tool page: open its card above.
What can you do with Qwen3.6 27B in Devin?
Qwen3.6 27B can help turn repository context into code fixes and working web features. Give the agent a concrete failure or acceptance criteria, then ask it to verify the resulting change.
Fix a reproducible bug
Give Qwen3.6 27B the failure, expected behavior, and relevant code. Request a focused patch and a regression check.
Build a web interface
Turn a written specification into a working interface. Include keyboard behavior, validation rules, and screen sizes in the request.
Explain repository behavior
Ask Qwen3.6 27B to trace a request through the relevant modules and identify where a proposed change belongs.
Write terminal utilities
Draft scripts for repetitive development tasks, with sample inputs, expected outputs, and explicit error handling.
Check a technical calculation
Supply the equations, units, and assumptions. Request a result with a dimensional check and a small numerical example.
Create a data visualization
Provide a small dataset and the comparison readers need to make. Request chart code with labeled axes and a written interpretation.
Prototype an interactive demo
Describe a small browser game or simulation, including controls and success conditions. Ask for a runnable implementation and checks for its core rules.
More on Qwen3.6 27B and Devin
Overview
Qwen3.6 27B in other agents
Questions about Qwen3.6 27B in Devin
Can Devin use Qwen3.6 27B?
Yes. Add Spiral to Devin as an MCP server, and Devin can find and run Qwen3.6 27B like any built-in tool.
How much does Qwen3.6 27B cost in Devin?
Up to $0.017055 per run at the default inputs. $0.416 per 1M input tokens, $4.16 per 1M output tokens. You pay from your Spiral balance as you go, with no subscription, and a failed run is not charged. Your Devin plan is separate.
Do I need an OpenRouter account to use Qwen3.6 27B in Devin?
No. Spiral runs Qwen3.6 27B on its own OpenRouter account. You sign in to Spiral once and never paste a key.
How long does it take to set up Qwen3.6 27B in Devin?
About two minutes: add a custom MCP, then sign in to Spiral and approve, then check it.
Does Devin wait while Qwen3.6 27B runs?
Qwen3.6 27B runs in the background. Devin starts the run, checks on it and gives you the text, json when it's ready.
What is Qwen3.6 27B good at?
Its clearest uses are coding tasks and STEM reasoning. Try it for debugging, building web interfaces, or working through technical problems with explicit constraints.
How much context can Qwen3.6 27B handle?
The native context length is 262,144 tokens. Qwen also documents extension to 1,010,000 tokens with additional configuration, so treat the larger figure as an extended limit.
Which settings should I try for precise coding?
Qwen recommends temperature 0.6, top_p 0.95, top_k 20, and min_p 0.0 for precise coding in thinking mode. Use that as a starting point and compare results on representative tasks.
How long can a Qwen3.6 27B response be?
Alibaba documents a maximum output length of 65,536 tokens in thinking mode. For a large implementation, request one tested change at a time to keep the result easier to review.
Where should I be cautious with Qwen3.6 27B?
Its reported SWE-bench results used an agent with bash and file-editing tools. That makes the scores evidence for a particular assisted workflow; review generated changes and test the behavior you need.
Start using Qwen3.6 27B in Devin
One Spiral account for every tool your agent needs. No subscription; you pay per run.
Paste this into Devin:
set up https://spiral.ai/SKILL.md


