Moduck

Sakana AI: Fugu

Multi-agent model for routing expert agents across complex analytical tasks

Model lab
Sakana AI
Model family
fugu

Model resources

Modalitiestext, image → text
Context window1,000,000
Maximum output—

Quickstart

Use your Moduck API Key, set the API base URL to https://api.moduck.ai/v1, and set the model parameter to sakana/fugu.

import os
import requests

base_url = "https://api.moduck.ai/v1"
api_key = os.environ["MODUCK_API_KEY"]
model_id = "sakana/fugu"

response = requests.post(
    f"{base_url}/chat/completions",
    headers={
        "Authorization": f"Bearer {api_key}",
        "Content-Type": "application/json",
    },
    json={
        "model": model_id,
        "messages": [{"role": "user", "content": "Hello!"}],
        "stream": False,
    },
    timeout=60,
)
response.raise_for_status()
print(response.json())

Use the familiar OpenAI-style chat request format to make model integration easier.

Fugu API Pricing

Price currently unavailable

Benchmarks

BenchmarkScoreMeasurementSource
SWE Bench Pro59—
Terminal Bench 2.180.2—
LiveCodeBench92.9—
LiveCodeBench Pro87.8—
Humanity’s Last Exam47.2—
CharXiv Reasoning85.1—
GPQA Diamond95.5—
SciCode60.1—
τ3 Banking21.7—
Long Context Reasoning74.7—
MRCRv286.6—
CTI-REALM67.5—

Performance

Performance trends

Throughput93.4 tok/sCurrent
Time to first token1.5 sCurrent
End-to-end latency8.0 sCurrent
Throughput
Current93.4 tok/s
Window average89.7 tok/s
Time to first token
Current1.5 s
Window average1.4 s
End-to-end latency
Current8.0 s
Window average7.5 s
Tail latency (P99)
Current13.3 s
Window average16.9 s
Prompt cache hit rate
Current31.2%
Window average22.6%
Error rate
Current0.0%
Window average0.2%

Availability

Uptime100%At least one provider responded
Availability100%Requests actually served
Error rate0%Share of failed requests
Served requests, hour by hour
24 hours agoNow

Fugu Compare models

ModelInput priceOutput priceContext windowCapabilities
FuguThis pagePrice currently unavailablePrice currently unavailable1,000,000Text output · Reasoning · Tool calling
Fugu UltraUSD 4 / 1M tokensUSD 24 / 1M tokens1,000,000Text output · Reasoning · Tool calling
Fugu MaxUSD 1.6 / 1M tokensUSD 4.8 / 1M tokens1,000,000Text output · Reasoning · Tool calling
Fugu Ultra v2USD 4 / 1M tokensUSD 24 / 1M tokens1,000,000Text output · Reasoning · Tool calling
DeepSeek V4.1 Flash 1017USD 0.028 / 1M tokens · Quantization fp4USD 0.4 / 1M tokens · Quantization fp41,000,000Text output · Reasoning · Tool calling

Model specifications

Input
text, image
Output
text
Context window
1,000,000
Maximum output
—
Release date
2026-06-15
Attachments
Yes
Reasoning
Yes
Tool calling
Yes
Structured output
Yes
Temperature setting
No
Open weights
No

Frequently asked questions

What is Fugu for?

Multi-agent model for routing expert agents across complex analytical tasks

What inputs and outputs does Fugu support?

Recorded inputs: text, images; outputs: text.

How is Fugu priced?

This page does not list a price for this model yet.

What specifications are recorded for Fugu?

Current model records list Context window 1,000,000 tokens.

What model parameter does the Quickstart use?

The example uses sakana/fugu. This value comes from the page identifier and is not confirmed as a callable Moduck public ID.

Can I call this model through the Moduck API now?

This page has no verified Moduck API call for this model. A Quickstart example or reserved ID does not establish access; check the models available to your account.