How to Use Qwen3.5 397B A17B in Codex
Use Qwen3.5 397B A17B for coding, reasoning, and tasks that involve several tool calls. Its 397B-parameter design activates 17B parameters per forward pass.
Prompts to try with Qwen3.5 397B A17B in Codex
Copy an example into Codex, or ask in your own words.
Use Qwen3.5 397B A17B to investigate the failing tests in this repository. Trace the failure to the relevant code, make a focused repair, and run the affected tests.
Use Qwen3.5 397B A17B to review the current diff for incorrect assumptions and missed edge cases. Report each finding with its file, line, and a concrete failure example.
Use Qwen3.5 397B A17B to compare the requirements in this project's documentation with the implementation. List contradictions and missing behavior before proposing changes.
Use Qwen3.5 397B A17B to plan a refactor of the most duplicated validation logic. Compare two approaches, preserve existing behavior, and identify the tests that would verify the refactor.
Use Qwen3.5 397B A17B with temperature 0.6, top_p 0.95, top_k 20, and min_p 0 to analyze the repository's retry logic. Find a reproducible edge case and add a regression test before changing the code.
How to set up Qwen3.5 397B A17B in Codex
Pick the way that suits you. Each takes about two minutes.
Use Spiral's MCP server
- Add the hosted server
codex mcp add spiral --url https://api.spiral.ai/mcp
- Sign in
Codex opens the browser: sign in to Spiral and click Authorize. No API key. codex mcp list then shows Spiral.
codex mcp login spiral
Use the Spiral skill
- Send Codex 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 Qwen3.5 397B A17B
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/qwen3.5-397b-a17b", "input": { "messages": [ { "role": "system" } ] } }'
Then paste one of the prompts above: Codex finds Qwen3.5 397B A17B on Spiral, shows the price and runs it.
Qwen3.5 397B A17B providers and prices
Spiral runs Qwen3.5 397B A17B through OpenRouter. Open a card for today's price, with every fee included.
Up to $0.015996 per run at the default inputs. $0.585 per 1M input tokens, $3.90 per 1M output tokens.
Settings you can ask for
Ask Codex 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 235929 | 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: Qwen3.5 397B A17B is on its tool page: open its card above.
What can you do with Qwen3.5 397B A17B in Codex?
Turn repository context into proposed repairs, implementation plans, and regression tests with Qwen3.5 397B A17B. Use concrete acceptance criteria and test the resulting changes.
Repair a failing test
Use Qwen3.5 397B A17B to trace a failure through the relevant code, propose a focused patch, and identify a regression test.
Compare implementation plans
Ask Qwen3.5 397B A17B to compare two technical approaches against explicit requirements, including edge cases and maintenance work.
Review long specifications
Give Qwen3.5 397B A17B specification text and request conflicting requirements, unanswered questions, and supporting passages.
Draft multilingual support replies
Ask Qwen3.5 397B A17B to draft replies in each customer's language using supplied policy text and a terminology guide.
Check mathematical solutions
Use Qwen3.5 397B A17B to examine a proposed solution, identify unsupported steps, and check the final result against the original conditions.
Plan tool actions
Give Qwen3.5 397B A17B a goal and available tool descriptions. Request an action sequence with prerequisites and a success check for each action.
Triage recurring technical reports
Use Qwen3.5 397B A17B to compare incoming error reports, group related symptoms, and propose the next diagnostic question.
More on Qwen3.5 397B A17B and Codex
Overview
Qwen3.5 397B A17B in other agents
Questions about Qwen3.5 397B A17B in Codex
Can Codex use Qwen3.5 397B A17B?
Yes. Add Spiral to Codex as an MCP server, and Codex can find and run Qwen3.5 397B A17B like any built-in tool.
How much does Qwen3.5 397B A17B cost in Codex?
Up to $0.019198 per run at the default inputs. $0.78 per 1M input tokens, $4.68 per 1M output tokens. You pay from your Spiral balance as you go, with no subscription, and a failed run is not charged. Your Codex plan is separate.
Do I need an OpenRouter account to use Qwen3.5 397B A17B in Codex?
No. Spiral runs Qwen3.5 397B A17B on its own OpenRouter account. You sign in to Spiral once and never paste a key.
How long does it take to set up Qwen3.5 397B A17B in Codex?
About two minutes: add the hosted server, then sign in.
Does Codex wait while Qwen3.5 397B A17B runs?
Qwen3.5 397B A17B runs in the background. Codex starts the run, checks on it and gives you the text, json when it's ready.
What is Qwen3.5 397B A17B good at?
It is a useful choice for code generation, reasoning, and work that requires several tool calls. Examples include proposing bug fixes, comparing technical approaches, and planning a sequence of actions.
What does 397B A17B mean?
The model has 397 billion parameters in total, with 17 billion activated per forward pass.
How much context does Qwen3.5 397B A17B support?
Its native context is 262,144 tokens, shared by input and output. Leave room for the response when supplying long material.
How long should I let it respond?
Qwen recommends an output budget of 32,768 tokens for most queries. That budget is a starting recommendation, not a requirement to produce a long answer or a stated maximum.
Which languages does Qwen3.5 support?
Qwen reports support for 201 languages and dialects. Specify the target language, audience, and terminology in your request.
Start using Qwen3.5 397B A17B in Codex
One Spiral account for every tool your agent needs. No subscription; you pay per run.
Paste this into Codex:
set up https://spiral.ai/SKILL.md


