Spiral gives your agent every tool with one account. Browse the tools →

Codex · LLMReleased 2026-04-03

How to Use Gemma 4 26B A4B in Codex

Gemma 4 26B A4B handles code generation, document summaries, and reasoning across long inputs. Unlike the dense Gemma 4 31B, it uses a mixture of experts with about 4 billion active parameters per token.

Prompts to try with Gemma 4 26B A4B in Codex

Copy an example into Codex, or ask in your own words.

Use Gemma 4 26B A4B to inspect the failing test and the code it exercises. Propose the smallest fix, explain the cause with file references, and add a regression test.

Gemma 4 26B A4BCodex
Try asking
  • Use Gemma 4 26B A4B to compare this branch's diff with the acceptance criteria in the project. List unmet requirements and point to the relevant code before proposing edits.

  • Use Gemma 4 26B A4B to review the request validation code. Suggest tests for empty values, boundary values, malformed input, and unexpected types using the project's test framework.

  • Use Gemma 4 26B A4B to explain how authentication flows through this repository. Cite the entry points and key functions, then list any behavior you could not confirm.

  • Use Gemma 4 26B A4B with thinking enabled to compare two refactoring approaches for the module I selected. Recommend one using the current interfaces and tests, and return a concise implementation plan.

How to set up Gemma 4 26B A4B in Codex

Pick the way that suits you. Each takes about two minutes.

Use Spiral's MCP server

  1. Add the hosted server
    codex mcp add spiral --url https://api.spiral.ai/mcp
  2. 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

  1. Send Codex this message

    Your agent reads Spiral's skill and sets itself up.

    set up https://spiral.ai/SKILL.md
  2. Sign in or paste a key

    Approve Spiral in the browser, or give the agent a key from the console.

Use the Spiral CLI

  1. 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

  1. Create an API key

    In the console, under API keys.

  2. Run Gemma 4 26B A4B

    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-4-26b-a4b-it",
        "input": {
          "messages": [
            {
              "role": "system"
            }
          ]
        }
      }'

Then paste one of the prompts above: Codex finds Gemma 4 26B A4B on Spiral, shows the price and runs it.

Gemma 4 26B A4B providers and prices

Spiral runs Gemma 4 26B A4B through OpenRouter. Open a card for today's price, with every fee included.

OpenRouterLLM
Google: Gemma 4 26B A4B
google/gemma-4-26b-a4b-it

Settings you can ask for

Ask Codex for any of these in your prompt.

SettingAllowed valuesDefaultDescription
Seeda whole number
Min p0 to 1
Top ka whole number from 0
Top p0 to 1
Logprobson or off
Max tokens1 to 235929

The most output tokens to generate.

Temperature0 to 2
Top logprobs0 to 20
Presence penalty-2 to 2
Frequency penalty-2 to 2
Repetition penalty0 to 2

Every API field of Google: Gemma 4 26B A4B is on its tool page: open its card above.

What can you do with Gemma 4 26B A4B in Codex?

Turn project code and written requirements into proposed fixes, tests, and explanations with Gemma 4 26B A4B. Give it the relevant files and ask for evidence that connects each change to the reported behavior.

Trace a software bug

Give Gemma 4 26B A4B the relevant code, error message, and expected behavior. Ask for a proposed fix and a regression test.

Generate tests from requirements

Turn written acceptance criteria into test cases covering normal behavior, boundaries, and invalid input. Review the proposed assertions before adding them.

Summarize lengthy reports

Condense supplied reports into decisions, supporting evidence, and unresolved questions. Request section references so a reviewer can check each conclusion.

Answer questions from a handbook

Provide the relevant handbook text and ask focused questions about its rules. Require a supporting passage and an explicit note when the answer is missing.

Draft multilingual replies

Prepare customer replies in a specified language using supplied facts and terminology. Ask for a separate list of wording that needs a fluent reviewer.

Compare proposals against requirements

Supply the proposals and evaluation criteria. Ask for a comparison that links each conclusion to evidence and marks missing information.

Explain results returned by tools

Use function calling to request relevant records, then have Gemma 4 26B A4B summarize the returned information into a report with record references.

Questions about Gemma 4 26B A4B in Codex

Can Codex use Gemma 4 26B A4B?

Yes. Add Spiral to Codex as an MCP server, and Codex can find and run Gemma 4 26B A4B like any built-in tool.

How much does Gemma 4 26B A4B cost in Codex?

Up to $0.001202 per run at the default inputs. $0.08775 per 1M input tokens, $0.2925 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 Gemma 4 26B A4B in Codex?

No. Spiral runs Gemma 4 26B A4B 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 26B A4B in Codex?

About two minutes: add the hosted server, then sign in.

Does Codex wait while Gemma 4 26B A4B runs?

Gemma 4 26B A4B runs in the background. Codex starts the run, checks on it and gives you the text, json when it's ready.

What is Gemma 4 26B A4B good for?

Use it to generate or review code, summarize supplied documents, and answer questions about long passages. These are practical starting points for its coding, text generation, and reasoning capabilities.

How much context can Gemma 4 26B A4B handle?

Google specifies a 256K-token context window. That describes context capacity; it does not establish a separate maximum answer length.

Which languages does Gemma 4 support?

Google lists more than 35 languages supported out of the box, with pretraining across more than 140 languages. Specify the language and audience when asking for a translation or rewrite.

Does Gemma 4 support thinking?

Yes. Thinking mode lets it reason before producing its final answer; try it for a debugging problem or a comparison with several constraints.

Which sampling settings should I start with?

Google recommends temperature 1.0, top_p 0.95, and top_k 64. Use that combination as a baseline when comparing results on your own tasks.

Get started

Start using Gemma 4 26B A4B in Codex

One Spiral account for every tool your agent needs. No subscription; you pay per run.

Paste this into Codex:

Set up Codex

set up https://spiral.ai/SKILL.md
Are you an AI agent? See setup options