How to Use Gemma 4 31B in OpenClaw
Gemma 4 31B suits code fixes, complex questions, and analysis of long documents. Google positions its dense architecture for answer quality, while Gemma 4 26B A4B prioritizes speed.
Prompts to try with Gemma 4 31B in OpenClaw
Copy an example into OpenClaw, or ask in your own words.
Each morning, use Gemma 4 31B to summarize the supplied support conversations from the previous day. Group recurring problems and include conversation identifiers for each finding.
Each week, use Gemma 4 31B to compare the supplied release announcements with last week's supplied summary. Report new changes and their source links in under 600 words.
For each new batch of supplied customer messages, use Gemma 4 31B to draft replies in each customer's language. Use only the supplied policy text and flag questions it does not answer.
Every Friday, use Gemma 4 31B to summarize the supplied research abstracts. List findings, disagreements, and questions for follow-up, preserving the publication identifiers.
For every supplied incident report, use Gemma 4 31B to extract the timeline, stated cause, corrective actions, and unresolved questions. Include evidence passages and leave unstated causes unknown.
How to set up Gemma 4 31B in OpenClaw
Pick the way that suits you. Each takes about two minutes.
Use the Spiral skill
- Send OpenClaw 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 Spiral's MCP server
- Add the server
On the machine that runs OpenClaw's Gateway:
openclaw mcp add spiral \ --url https://api.spiral.ai/mcp \ --transport streamable-http \ --auth oauth
- Sign in
In Settings › MCP, select Sign in on the Spiral server, sign in to Spiral and click Authorize. No API key.
openclaw mcp login spiral
- Reload and check
A running Gateway may keep its old tool list: restart it, or Save & Publish in the Control UI.
openclaw mcp reload openclaw mcp doctor spiral --probe
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 Gemma 4 31B
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": "google/gemma-4-31b-it", "input": { "messages": [ { "role": "system" } ] } }'
Then paste one of the prompts above: OpenClaw finds Gemma 4 31B on Spiral, shows the price and runs it.
Gemma 4 31B providers and prices
Spiral runs Gemma 4 31B through OpenRouter. Open a card for today's price, with every fee included.
Up to $0.00214 per run at the default inputs. $0.26 per 1M input tokens, $0.52 per 1M output tokens.
Settings you can ask for
Ask OpenClaw 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 16384 | 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 Google: Gemma 4 31B is on its tool page: open its card above.
What can you do with Gemma 4 31B in OpenClaw?
Turn recurring batches of supplied text into summaries, extracted records, and draft replies with Gemma 4 31B. Define the evidence and reporting rules so each scheduled run produces results you can compare.
Diagnose a code defect
Ask Gemma 4 31B to trace a failing path, propose a small correction, and suggest a regression check.
Compare long documents
Use Gemma 4 31B to compare requirements, identify contradictions, and attach supporting passages to each finding.
Extract structured records
Ask Gemma 4 31B to turn supplied text into JSON records, using null when information is absent.
Draft multilingual responses
Use Gemma 4 31B to draft replies in a requested language while preserving names and approved terminology.
Summarize research
Ask Gemma 4 31B to condense supplied research into findings, disagreements, and unanswered questions.
Check a plan against constraints
Give Gemma 4 31B a proposed plan and explicit constraints, then request conflicts and a revised sequence.
Answer questions from reference text
Use Gemma 4 31B to draft answers grounded in supplied reference material, with evidence for each conclusion.
More on Gemma 4 31B and OpenClaw
Overview
Gemma 4 31B in other agents
Questions about Gemma 4 31B in OpenClaw
Can OpenClaw use Gemma 4 31B?
Yes. Add Spiral to OpenClaw as an MCP server, and OpenClaw can find and run Gemma 4 31B like any built-in tool.
How much does Gemma 4 31B cost in OpenClaw?
Up to $0.00214 per run at the default inputs. $0.26 per 1M input tokens, $0.52 per 1M output tokens. You pay from your Spiral balance as you go, with no subscription, and a failed run is not charged. Your OpenClaw plan is separate.
Do I need an OpenRouter account to use Gemma 4 31B in OpenClaw?
No. Spiral runs Gemma 4 31B on its own OpenRouter account. You sign in to Spiral once and never paste a key.
How long does it take to set up Gemma 4 31B in OpenClaw?
About two minutes: add the server, then sign in, then reload and check.
Does OpenClaw wait while Gemma 4 31B runs?
Gemma 4 31B runs in the background. OpenClaw starts the run, checks on it and gives you the text, json when it's ready.
What is Gemma 4 31B useful for?
Use it for code fixes, questions that require reasoning, summaries of long documents, and extracting information into JSON. Google positions the dense 31B model for answer quality.
How much context can Gemma 4 31B handle?
Google specifies a model context window of up to 256K tokens. That describes context capacity, not the length of a single answer.
Can Gemma 4 31B work in languages other than English?
Yes. Google describes Gemma 4 as supporting 140 languages; name the target language and explain which terms, names, or quotations should stay unchanged.
Does Gemma 4 31B have a thinking mode?
Yes. Thinking mode produces reasoning before the final answer; try it for code diagnosis, constraint checking, or questions with several dependent steps.
Which generation settings should I start with?
Google recommends temperature 1.0, nucleus sampling at 0.95, and a top-k value of 64 across use cases. Use those as a baseline before comparing changes on examples from your task.
Start using Gemma 4 31B in OpenClaw
One Spiral account for every tool your agent needs. No subscription; you pay per run.
Paste this into OpenClaw:
set up https://spiral.ai/SKILL.md


