[Integrations](https://spiral.ai/integrations) › [Factory Droid](https://spiral.ai/integrations/factory-droid) › [Gemini 4 Carbon Batch](https://spiral.ai/integrations/factory-droid/gemini-4-carbon-batch)

# How to Use Gemini 4 Carbon Batch in Factory Droid

Use Gemini 4 Carbon Batch for difficult debugging, architectural reviews, and scientific problem solving. Compared with Gemini 4 Argon, Carbon spends more tokens exploring possibilities and checking its work.

|  |  |
| --- | --- |
| Model | Gemini 4 Carbon Batch |
| Available through | OpenRouter |
| Category | LLM |
| Price | Up to $0.02667 per run at the default inputs. $1.30 per 1M input tokens, $6.50 per 1M output tokens. |
| Returns | text, json |
| Runs | In the background |
| Setup | About 2 minutes, no API key |

## Prompts to try with Gemini 4 Carbon Batch in Factory Droid

- Use Gemini 4 Carbon Batch with high reasoning to investigate the intermittent checkout failure in this repository. Trace the request across services, reproduce the failure, make the smallest justified fix, and run a regression check.
- Use Gemini 4 Carbon Batch to review this project's package migration plan. Map affected imports and interfaces, identify compatibility risks, and produce an ordered plan with a verification step for each change.
- Use Gemini 4 Carbon Batch with high reasoning to review the architecture in this repository. Compare the current design with two alternatives against the documented traffic and recovery requirements, and cite the code supporting each concern.
- Use Gemini 4 Carbon Batch to implement the browser application described in the project specification. Check each acceptance criterion, run the relevant build checks, and report any behavior you could not verify.
- Use Gemini 4 Carbon Batch to audit the numerical calculations in this project's simulation. Check units, boundary conditions, and limiting cases, then propose targeted checks for any suspected error.

## How to set up Gemini 4 Carbon Batch in Factory Droid

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

### Use Spiral's MCP server

1. **Add the server**: Droid saves it to your user config, `~/.factory/mcp.json`, so it works in every project.

```bash
droid mcp add spiral https://api.spiral.ai/mcp --type http
```

2. **Sign in**: Start `droid`, type `/mcp`, select spiral and authenticate it. Sign in to Spiral in the browser and click Authorize. No API key.

3. **Check it**: Spiral shows as connected. needs authentication means the sign-in did not finish: run `/mcp` again.

```bash
droid mcp list
```

### Use the Spiral skill

1. **Send Factory Droid 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.

```bash
npx -y spiral-cli@latest init
```

### Call the Spiral API

1. **Create an API key**: In the console, under API keys.

2. **Run Gemini 4 Carbon Batch**: One request; the result comes back by polling or a webhook.

```bash
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/gemini-4-carbon:batch",
    "input": {
      "messages": [
        {
          "role": "system"
        }
      ]
    }
  }'
```

Then paste one of the prompts above: Factory Droid finds Gemini 4 Carbon Batch on Spiral, shows the price and runs it.

Setup guide in the docs: https://spiral.ai/docs/quickstarts/factory-droid

## Gemini 4 Carbon Batch tools and prices

- [Google: Gemini 4 Carbon (batch)](https://spiral.ai/tools/openrouter/google/gemini-4-carbon%3Abatch) (OpenRouter): Up to $0.02667 per run at the default inputs. $1.30 per 1M input tokens, $6.50 per 1M output tokens.

## Settings you can ask for

| Setting | Allowed values | Default | Description |
| --- | --- | --- | --- |
| seed |  | — |  |
| stop |  | — |  |
| tools |  | — |  |
| top_p | 0 to 1 | — |  |
| reasoning |  | — | Reasoning effort; reasoning tokens count toward max_tokens. |
| max_tokens | 1 to 943718 | — | The most output tokens to generate. On a reasoning model the reasoning tokens count toward it: a cap too low for the effort can end with no answer (finish_reason length). |

## What can you do with Gemini 4 Carbon Batch in Factory Droid?

Investigate failures across a codebase and carry a proposed fix through verification with Gemini 4 Carbon. Give the agent a reproducible problem, project constraints, and acceptance criteria.

### Investigate a stubborn bug

Trace a failure across components, compare possible causes, and propose a fix with a regression check.

### Review a migration plan

Examine a proposed package migration for dependency conflicts, compatibility risks, and missing verification steps.

### Challenge an architecture decision

Compare designs against stated constraints and identify the assumptions that could change the recommendation.

### Build a browser application

Turn a written specification into application code, then check the implementation against acceptance criteria.

### Analyze experimental results

Compare competing explanations for supplied measurements and propose experiments that could distinguish them.

### Review a mathematical argument

Check a proposed proof for missing assumptions, invalid steps, and counterexamples.

### Prepare an analytical report

Turn supplied evidence and calculations into a report with findings, unresolved questions, and supporting references.

### Audit a spreadsheet model

Review supplied formulas and assumptions for inconsistent units, missing dependencies, and sensitivity to uncertain inputs.

## Overview

- [OpenRouter](https://spiral.ai/providers/openrouter)
- [All Factory Droid guides](https://spiral.ai/integrations/factory-droid)

## Gemini 4 Carbon Batch in other agents

- [Amp](https://spiral.ai/integrations/amp/gemini-4-carbon-batch)
- [ChatGPT](https://spiral.ai/integrations/chatgpt/gemini-4-carbon-batch)
- [Claude](https://spiral.ai/integrations/claude-desktop/gemini-4-carbon-batch)
- [Claude Code](https://spiral.ai/integrations/claude-code/gemini-4-carbon-batch)
- [Cline](https://spiral.ai/integrations/cline/gemini-4-carbon-batch)
- [Codex](https://spiral.ai/integrations/codex/gemini-4-carbon-batch)
- [Cursor](https://spiral.ai/integrations/cursor/gemini-4-carbon-batch)
- [Devin](https://spiral.ai/integrations/devin/gemini-4-carbon-batch)
- [dots](https://spiral.ai/integrations/dots/gemini-4-carbon-batch)
- [Gemini](https://spiral.ai/integrations/gemini/gemini-4-carbon-batch)
- [Gemini CLI](https://spiral.ai/integrations/gemini-cli/gemini-4-carbon-batch)
- [Google Antigravity](https://spiral.ai/integrations/antigravity/gemini-4-carbon-batch)
- [goose](https://spiral.ai/integrations/goose/gemini-4-carbon-batch)
- [Grok](https://spiral.ai/integrations/grok/gemini-4-carbon-batch)
- [Grok Build](https://spiral.ai/integrations/grok-build/gemini-4-carbon-batch)
- [Hermes](https://spiral.ai/integrations/hermes/gemini-4-carbon-batch)
- [Instinct](https://spiral.ai/integrations/instinct/gemini-4-carbon-batch)
- [Manus](https://spiral.ai/integrations/manus/gemini-4-carbon-batch)
- [Muse](https://spiral.ai/integrations/muse/gemini-4-carbon-batch)
- [Muse Code](https://spiral.ai/integrations/muse-code/gemini-4-carbon-batch)
- [n8n](https://spiral.ai/integrations/n8n/gemini-4-carbon-batch)
- [Notion](https://spiral.ai/integrations/notion/gemini-4-carbon-batch)
- [OpenClaw](https://spiral.ai/integrations/openclaw/gemini-4-carbon-batch)
- [opencode](https://spiral.ai/integrations/opencode/gemini-4-carbon-batch)
- [Perplexity](https://spiral.ai/integrations/perplexity/gemini-4-carbon-batch)
- [Raycast](https://spiral.ai/integrations/raycast/gemini-4-carbon-batch)
- [VS Code](https://spiral.ai/integrations/vscode/gemini-4-carbon-batch)
- [Windsurf](https://spiral.ai/integrations/windsurf/gemini-4-carbon-batch)
- [Zapier](https://spiral.ai/integrations/zapier/gemini-4-carbon-batch)

## More Factory Droid guides

- [Gemini 4 Carbon](https://spiral.ai/integrations/factory-droid/gemini-4-carbon)
- [Gemini 4 Hydrogen](https://spiral.ai/integrations/factory-droid/gemini-4-hydrogen)
- [Gemini 4 Argon](https://spiral.ai/integrations/factory-droid/gemini-4-argon)
- [Gemini 4 Hydrogen Vertex AI](https://spiral.ai/integrations/factory-droid/gemini-4-hydrogen-vertex-ai)
- [Gemini 3.8 Flash](https://spiral.ai/integrations/factory-droid/gemini-3-8-flash)
- [Gemini 4 Neon](https://spiral.ai/integrations/factory-droid/gemini-4-neon)

## Questions about Gemini 4 Carbon Batch in Factory Droid

**Can Factory Droid use Gemini 4 Carbon Batch?**

Yes. Add Spiral to Factory Droid as an MCP server, and Factory Droid can find and run Gemini 4 Carbon Batch like any built-in tool.

**How much does Gemini 4 Carbon Batch cost in Factory Droid?**

Up to $0.02667 per run at the default inputs. $1.30 per 1M input tokens, $6.50 per 1M output tokens. You pay from your Spiral balance as you go, with no subscription, and a failed run is not charged. Your Factory Droid plan is separate.

**Do I need an OpenRouter account to use Gemini 4 Carbon Batch in Factory Droid?**

No. Spiral runs Gemini 4 Carbon Batch on its own OpenRouter account. You sign in to Spiral once and never paste a key.

**How long does it take to set up Gemini 4 Carbon Batch in Factory Droid?**

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

**Does Factory Droid wait while Gemini 4 Carbon Batch runs?**

Gemini 4 Carbon Batch runs in the background. Factory Droid starts the run, checks on it and gives you the text, json when it's ready.

**What is Gemini 4 Carbon best used for?**

Use it for difficult debugging, architectural reviews, scientific problems, and detailed analytical reports. Give it a concrete question and a way to check the result.

**How does Gemini 4 Carbon differ from Gemini 4 Argon?**

Carbon spends more tokens exploring possibilities and running verification loops. That makes it a useful choice when a task needs deeper investigation and repeated checks.

**How much text can Gemini 4 Carbon handle?**

Google documents a shared 1,048,576-token window for input, thinking, and output. Long decode supports generation beyond 64K tokens within that space, so a larger input leaves less room for reasoning and the answer.

**Can Gemini 4 Carbon invent facts?**

Yes. Google lists hallucinations as a limitation and gives a March 2026 knowledge cutoff. Supply current evidence and ask it to distinguish supported conclusions from assumptions.

**Which settings matter for Gemini 4 Carbon?**

Reasoning effort is the useful control: medium is the default, and high allows Carbon's deepest reasoning. Google says temperature and sampling controls are ignored, so adjust the task instructions and reasoning effort instead.

