Last verified
MULTI-AGENT MODELOPENAI-COMPATIBLE APIFUGU ULTRAPAY AS YOU GONO EU/EEA

Sakana Fugu API Pricing

Sakana AI describes Sakana Fugu as a multi-agent system delivered as one model API. Fugu dynamically orchestrates a pool of expert models for coding, reasoning, research, security analysis, and other complex multi-step work.

Priced models
2
2 GA
Fugu Ultra input
$5/M
Fugu Ultra
Fugu Ultra output
$30/M
Fugu Ultra
Context threshold
272K
higher rates above
API style
OpenAI
compatible endpoint
EU/EEA
not yet
official availability note
§ 01 / LINEUP

The full roster.

Side-by-side →
§ 02 / SHELF

All side-by-side.

Methodology →
Model Input /M Output /M Cached Context Max output Vision Tools Tier
Fugu Ultra FLAGSHIP $5.00 $30.00 $0.50−90% 272K Active
Sakana Namazu $0.95 $4.00 $0.15−84% 256K Active

Sakana Fugu pricing notes.

AI//COST tracks Fugu Ultra because Sakana publishes a fixed token row for fugu-ultra-20260615. Regular Fugu remains variable-rate because its price depends on the active underlying model tier.

AUG 23 · 2026
Sakana Namazu enters the catalogue — the lab's first in-house single model, Japanese-specialised, at $0.95/$4.00 with $0.15 cached; built on Kimi K2.6 and priced at that base model's own rate
RELEASE
JUN 15 - 2026
fugu-ultra-20260615 dated model ID appears in Sakana pricing
MODEL
JUN 28 - 2026
Fugu Ultra added to AI//COST from the official Sakana Fugu page
VERIFY
2026
Sakana Fugu listed as unavailable in the EU/EEA while compliance work continues
ACCESS
§ 04 / ACCESS

Where to get it.

Methodology →
§ 04 / BEST FOR

Which Sakana AI for what.

More scenarios →
If you need maximum Sakana Fugu quality...
Fugu Ultra
Profile ->
If you need balanced latency and cost...
Fugu
Vendor ->
If the work is in Japanese...
Sakana Namazu
Profile ->
If your workload is inside EU/EEA...
Wait / use another provider
Compare ->
§ 06 / BACKGROUND

The company behind it.

sakana.ai →

Sakana Fugu is positioned as "Multi-Agent System as a Model": one OpenAI-compatible model API that dynamically coordinates a pool of expert models for complex, multi-step tasks. The public page emphasizes coding, reasoning, Kaggle competitions, paper reproduction, cybersecurity analysis, and literature or patent investigations.

AI//COST tracks two token-priced rows here. Fugu Ultra carries a fixed pay-as-you-go card, now served under the fugu-ultra alias at v1.1 (the dated fugu-ultra-20260615 id is the vendor's own name for v1.0, and both bill identically). Sakana Namazu, added in August 2026, is the lab's first in-house single model rather than an orchestrator: Japanese-specialised, built on Kimi K2.6, and priced at $0.95/$4.00 — the same rates Moonshot charges for that base model. It is pay-as-you-go only, outside the monthly plans, and its built-in tools bill separately at $7.00 per 1,000 search calls and $0.12 per code-execution hour. Regular Fugu remains variable-rate: Sakana says a single active agent is billed at that underlying model rate, and multiple active agents are billed as one rate based on the top-tier model involved.

§ 07 / COMPETITORS

Other frontier labs.

All providers →
FRONTIER API

OpenAI

GPT-5.6 Sol used to carry the identical $5/$30/$0.50 row; since its August 2026 cut to $4/$20/$0.40 it undercuts Fugu Ultra outright, and without Sakana's multi-agent orchestration on top.

GPT family
AGENT MODEL

Anthropic

Claude Fable 5 and Mythos 5 are key comparators for high-end agentic work and long-context coding/research tasks.

Claude family
LONG CONTEXT

Google

Gemini provides lower-cost long-context alternatives when a single-model workflow is enough.

Gemini family
OPEN-WEIGHT GIANT

Alibaba (Qwen)

The Qwen (Tongyi Qianwen) model family from Alibaba Cloud's Tongyi Lab. First released in 2023, Qwen is the most-downloaded open-weight model family in the world — most tiers ship under Apache 2.0 on Hugging Face and ModelScope, while the proprietary Max tier is served through Alibaba Cloud's Model Studio.

33 models · from $0.03/M
GLM ARCHITECT

Zhipu (Z.ai / GLM)

The Beijing lab behind the GLM models, also operating internationally as Z.ai. Spun out of Tsinghua University in 2019, Zhipu ships strong coding/agentic models, multiple free tiers, and open weights — and was the first of China's "Six Tigers" to pursue an IPO. It also sits on the US Entity List.

16 models · from $0.07/M
CHINA CONSUMER LEADER

ByteDance (Doubao)

The Doubao model family from ByteDance — parent of TikTok and Douyin. Served through the Volcano Ark (火山方舟) platform on ByteDance's Volcano Engine cloud, Doubao powers China's most-used consumer AI app and undercuts most frontier labs on price.

14 models · from $0.0223/M

Frequently asked.

Practical notes on fixed Fugu Ultra prices, variable Fugu billing, and availability limits.

Q · 01 What Fugu prices are tracked here? +
AI//COST tracks the fixed Fugu Ultra token row from sakana.ai/fugu, verified 2026-06-28.
Q · 02 Why is regular Fugu not shown as a normal fixed price? +
Sakana describes regular Fugu as variable: one active agent uses that underlying model rate, while multiple active agents are charged one rate based on the top-tier model involved.
Q · 03 Does Fugu Ultra have a long-context surcharge? +
Yes. Sakana lists $10/M input, $45/M output, and $1.00/M cached input when context exceeds 272K.
Q · 04 Is Sakana Fugu available in the EU/EEA? +
No. The official page says Fugu is not yet available in the EU/EEA while Sakana works toward GDPR and EU-specific regulations.