[Integrations](https://spiral.ai/integrations) › [Claude Code](https://spiral.ai/integrations/claude-code) › [Fugu Max](https://spiral.ai/integrations/claude-code/fugu-max)

# How to Use Fugu Max in Claude Code

Fugu Max coordinates expert models for coding, reasoning, and document analysis. It targets efficiency alongside capability, while Fugu Ultra v2 prioritizes maximum answer quality.

|  |  |
| --- | --- |
| Model | Fugu Max |
| Available through | OpenRouter |
| Category | LLM |
| Price | Up to $0.03204 per run at the default inputs. $2.60 per 1M input tokens, $7.80 per 1M output tokens. |
| Returns | text, json |
| Runs | In the background |
| Setup | About 2 minutes, no API key |

## Prompts to try with Fugu Max in Claude Code

- Use Fugu Max to review the current diff and relevant callers. Report only findings with a concrete failure path, affected location, and regression test.
- Use Fugu Max to investigate the failing test output in this project. Rank three plausible causes, run checks that distinguish them, and propose the smallest supported fix.
- Use Fugu Max with max reasoning to compare the two implementation proposals in this repository. Check each against the existing interfaces and tests, then recommend one with explicit tradeoffs.
- Use Fugu Max to trace how this project handles retries and duplicate requests. Identify where repeated execution could change state twice and propose focused regression tests.
- Use Fugu Max to analyze the project's lint and test findings. Return JSON objects containing location, evidence, proposed fix, and verification step; leave unsupported conclusions out.

## How to set up Fugu Max in Claude Code

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

### Use Spiral's MCP server

1. **Add the hosted server**: This adds Spiral to the current project. Add `--scope user` to use it in every project.

```bash
claude mcp add --transport http spiral https://api.spiral.ai/mcp
```

2. **Sign in**: Start `claude`, type `/mcp`, choose spiral and Authenticate. Sign in to Spiral in the browser and click Authorize. No API key.

### Use the Spiral skill

1. **Send Claude Code 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 Fugu Max**: 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": "sakana/fugu-max",
    "input": {
      "messages": [
        {
          "role": "system"
        }
      ]
    }
  }'
```

Then paste one of the prompts above: Claude Code finds Fugu Max on Spiral, shows the price and runs it.

Setup guide in the docs: https://spiral.ai/docs/quickstarts/claude-code

## Fugu Max tools and prices

- [Sakana: Fugu Max](https://spiral.ai/tools/openrouter/sakana/fugu-max) (OpenRouter): Up to $0.03204 per run at the default inputs. $2.60 per 1M input tokens, $7.80 per 1M output tokens.

## Settings you can ask for

| Setting | Allowed values | Default | Description |
| --- | --- | --- | --- |
| tools |  | — |  |
| reasoning |  | — | Reasoning effort; reasoning tokens count toward max_tokens. |
| tool_choice |  | — |  |
| reasoning_effort | high, xhigh, max | high | How much the model reasons. |

## What can you do with Fugu Max in Claude Code?

Fugu Max can help investigate failures, review changes, and reason across project code. Give it a concrete acceptance test and ask for evidence behind each proposed fix.

### Review a proposed code change

Ask Fugu Max to trace a patch through its callers, identify concrete failure paths, and propose a regression test for each finding.

### Investigate a failing build

Give Fugu Max the error output and relevant code. Request ranked explanations, checks that distinguish them, and the smallest proposed fix.

### Compare technical proposals

Have Fugu Max compare competing designs against explicit requirements, identify conflicting assumptions, and explain which tradeoffs affect the decision.

### Reconcile long documents

Ask Fugu Max to compare PDF specifications, cite the pages behind each difference, and separate confirmed changes from unclear wording.

### Check a quantitative argument

Use Fugu Max to examine a calculation or technical derivation, identify unsupported assumptions, and check the conclusion against boundary cases.

### Turn evidence into structured findings

Ask Fugu Max to organize supplied reports into JSON findings with evidence, severity, and open questions. Define the schema before requesting the analysis.

## Overview

- [OpenRouter](https://spiral.ai/providers/openrouter)
- [All Claude Code guides](https://spiral.ai/integrations/claude-code)

## Fugu Max in other agents

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

## More Claude Code guides

- [Ember-1](https://spiral.ai/integrations/claude-code/ember-1)
- [Fugu Ultra](https://spiral.ai/integrations/claude-code/fugu-ultra)
- [Elorian Glimpse](https://spiral.ai/integrations/claude-code/elorian-glimpse)
- [Fugu Ultra v2](https://spiral.ai/integrations/claude-code/fugu-ultra-v2)
- [ERNIE-Light-DT-1.0](https://spiral.ai/integrations/claude-code/ernie-light-dt-1-0)
- [GLM 4.5 Air](https://spiral.ai/integrations/claude-code/glm-4-5-air)

## Questions about Fugu Max in Claude Code

**Can Claude Code use Fugu Max?**

Yes. Add Spiral to Claude Code as an MCP server, and Claude Code can find and run Fugu Max like any built-in tool.

**How much does Fugu Max cost in Claude Code?**

Up to $0.03204 per run at the default inputs. $2.60 per 1M input tokens, $7.80 per 1M output tokens. You pay from your Spiral balance as you go, with no subscription, and a failed run is not charged. Your Claude Code plan is separate.

**Do I need an OpenRouter account to use Fugu Max in Claude Code?**

No. Spiral runs Fugu Max on its own OpenRouter account. You sign in to Spiral once and never paste a key.

**How long does it take to set up Fugu Max in Claude Code?**

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

**Does Claude Code wait while Fugu Max runs?**

Fugu Max runs in the background. Claude Code starts the run, checks on it and gives you the text, json when it's ready.

**What is Fugu Max best suited for?**

Consider it for coding, multistep reasoning, and document analysis. Sakana reports strong benchmark results across terminal tasks, reasoning, documents, automation, and software engineering.

**How does Fugu Max differ from Fugu Ultra v2?**

Fugu Max targets capability and efficiency together. Sakana positions Fugu Ultra v2 for workloads where maximum answer quality takes priority.

**How much text can Fugu Max handle?**

Its context window is 1,000,000 tokens, and it supports up to 128,000 completion tokens. Keep requested reports focused on the sections you actually need.

**Can Fugu Max analyze PDFs?**

Yes, it accepts PDFs and returns text. Try asking for a comparison table with page references and a separate list of unresolved questions.

**Does max reasoning go further than xhigh?**

No. Sakana documents xhigh and max as aliases for the same reasoning effort, so switching between them does not add another reasoning level.

