The confirmed way to access Sakana Fugu is through the Sakana console: sign in at console.sakana.ai with Google or email, then use the dashboard to find your API keys, available models, pricing, and billing options. There is no separate waitlist form on the official release page, and there is no standalone free tier today. If you want the closest thing to a trial, verify the reported “free second month” launch promo in the console. For a deeper API walkthrough, see our companion guide on how to use the Sakana Fugu API.
What Fugu is before you sign up
Fugu is not a single language model in the usual sense. Sakana presents it as a multi-agent orchestration system behind one OpenAI-compatible API. The release headline is “One Model to Command Them All.”
Under the hood, Fugu specializes in:
- delegating work across agents
- coordinating agent communication
- synthesizing outputs
- dynamically choosing which model or agent should handle each part of a task
- recursively invoking instances of itself when needed
There are two current variants:
| Variant | Best for |
|---|---|
fugu |
Lower-latency everyday work, coding, code review, chatbots, and interactive services |
fugu-ultra |
Maximum answer quality, AI research, paper reproduction, cybersecurity analysis, literature review, and patent investigation |
Early beta coverage often referred to the smaller variant as “Fugu Mini.” In the console, look for the current names: Fugu and Fugu Ultra.
Both variants use the same login, API key flow, and endpoint. You do not sign up for separate accounts per model. You authenticate once, then choose the model ID per request.
Access checklist
Use this flow to check whether you can access Fugu:
- Open console.sakana.ai.
- Sign in with Google or email.
- Check whether the dashboard lets you access:
- API keys
- model list
- pricing or billing
- usage information
- Copy your API key.
- Copy the base URL from the console.
- Confirm the exact model IDs shown in your account.
Do not guess the API host. As of 2026-06-22, Sakana has not published the API base URL on a public page. Treat the console as the source of truth.
If you see a hardcoded Fugu API host in a forum or third-party article, verify it against your own dashboard before using it in code.
This access pattern is similar to other gated model previews: sign in, find keys, then call the API. For comparison, see our walkthrough on how to access Claude Fable 5.
Beta versus general availability
Fugu previously ran a closed beta with roughly 500 users starting around April 24–25, 2026. That was an invite-style cohort, not open self-serve access.
The June 22 release introduced self-serve subscription tiers and a pay-as-you-go plan. Self-serve pricing usually indicates general availability, but you should still verify the live sign-up flow yourself.
Use this decision tree:
| What you see in the console | What it likely means |
|---|---|
| You can create an account, add billing, and generate an API key | You have access |
| You can sign in but cannot create keys | Your account may still be gated |
| You see a waitlist or “access pending” state | Access is not open for your account |
| Billing or sign-up is blocked by location | A regional restriction may apply |
[VERIFY 2026-06-22: whether GA self-serve sign-up is open to all, or still gated.]
Region availability
There are reports of an EU and EEA availability restriction at launch. This has not been confirmed against an official statement.
[VERIFY 2026-06-22: reported EU/EEA availability restriction.]
If you are in the EU or EEA and the console blocks sign-up or billing, region availability is the likely reason. Check Sakana’s terms and the console behavior before planning a production rollout.
Is Sakana Fugu free?
No. There is no standalone free tier today.
The release page confirms the pricing structure in concept:
- subscription tiers for everyday usage
- pay-as-you-go for heavier or enterprise workloads
However, specific dollar amounts circulating online come from JS-rendered or secondary sources, not the static release page. Confirm all prices inside your own console before budgeting.
Reported pricing, verify live as of 2026-06-22: secondary sources cite three subscription tiers at $20, $100, and $200 per month covering both models, a launch promo offering a free second month if you subscribe before the end of July 2026, and pay-as-you-go rates around $5 input, $30 output, and $0.50 cached per 1M tokens, with a surcharge above 272K context. The base “Fugu” variant is reportedly passthrough-billed at the standard rate of the underlying model it calls. None of these figures appear on the official release page. Confirm them inside console.sakana.ai before budgeting.
The reported “free second month” promo is not a free tier. It is a discount on a paid subscription.
For the full cost breakdown and comparison between subscription tiers and pay-as-you-go, see our Sakana Fugu pricing guide.
If your goal is free or low-cost model access, compare providers before committing. Our roundup of the best OpenRouter alternatives covers where free and low-cost tiers exist across the market. The same pattern appears in how to use Claude Fable 5 for free: usually you are looking at trials and promos, not perpetual free access.
What you get after access
Once your account is active, you get an OpenAI-compatible API endpoint.
That means you can reuse existing OpenAI SDK-based code by changing:
api_keybase_urlmodel
You do not need a dedicated Sakana SDK for basic chat completions.
Reported model IDs include:
fugu
fugu-ultra
fugu-ultra-20260615
Confirm the exact strings inside your console before shipping code.
Minimal Python example
Install the OpenAI SDK if you do not already have it:
pip install openai
Then call Fugu using the base URL from your Sakana console:
from openai import OpenAI
client = OpenAI(
api_key="YOUR_SAKANA_API_KEY",
base_url="<YOUR_FUGU_BASE_URL_FROM_CONSOLE>",
)
response = client.chat.completions.create(
model="fugu", # verify the exact model ID in your console
messages=[
{
"role": "user",
"content": "Summarize the tradeoffs of multi-agent orchestration."
}
],
)
print(response.choices[0].message.content)
To use Fugu Ultra, change the model:
response = client.chat.completions.create(
model="fugu-ultra",
messages=[
{
"role": "user",
"content": "Review this architecture for reliability risks."
}
],
)
The call shape follows the OpenAI chat completions format. The Fugu-specific parts are the base URL, key, and model ID.
Suggested implementation workflow
When you get console access, avoid wiring Fugu directly into production first. Use this sequence:
- Generate an API key in the Sakana console.
- Store the key in your secrets manager or local
.env. - Copy the base URL from the console.
- Make a basic chat completion call.
- Log the raw response shape.
- Test both
fuguandfugu-ultra. - Compare latency, answer quality, and cost.
- Add retries and timeout handling.
- Add request and response logging with sensitive data redacted.
- Only then connect the model to your app flow.
Example .env layout:
SAKANA_API_KEY=your_key_here
SAKANA_BASE_URL=https://your-console-provided-base-url
SAKANA_MODEL=fugu
Example Python configuration:
import os
from openai import OpenAI
client = OpenAI(
api_key=os.environ["SAKANA_API_KEY"],
base_url=os.environ["SAKANA_BASE_URL"],
)
model = os.getenv("SAKANA_MODEL", "fugu")
Governance note
Sakana says agents in the pool are swappable, teams can opt specific agents out for data or compliance reasons, and Fugu dynamically routes around provider restrictions.
If your app handles regulated or sensitive data, confirm these controls before production use:
- which agents can receive your prompts
- whether any providers must be excluded
- how data is routed
- what retention policies apply
- whether regional restrictions affect your workload
How this fits your Apidog workflow
After you have a key and base URL, use Apidog to test and document your Fugu requests.
Because Fugu uses an OpenAI-compatible request format, you can create a reusable request in Apidog with:
- method:
POST - endpoint: your Fugu base URL plus the chat completions path shown in your console
- auth: Bearer token
- body: OpenAI-compatible chat completion JSON
A typical request body looks like this:
{
"model": "fugu",
"messages": [
{
"role": "user",
"content": "Explain how multi-agent orchestration changes API design."
}
]
}
Using Apidog helps you:
- save the request instead of keeping one-off curl commands
- share the request with teammates
- document request and response schemas
- test different models in one workspace
- mock the endpoint while waiting for access
- compare Fugu against other OpenAI-compatible providers
Frequently asked questions
Where do I sign up for Sakana Fugu?
Sign up at console.sakana.ai using Google or email. There is no separate waitlist form linked from the official release page as of 2026-06-22. After sign-in, check the console dashboard for keys, model IDs, pricing, and billing.
Is Sakana Fugu free to use?
No. There is no standalone free tier today. The cheapest entry is the lowest reported subscription tier, and the closest thing to a trial is the reported “free second month” launch promo. Confirm every price in the console. Our Sakana Fugu pricing guide tracks the details.
Is the beta still running, or is Fugu generally available?
The closed beta ran with roughly 500 users starting in late April 2026. The June 22 release adds self-serve subscription tiers, which usually suggests general availability. We have not confirmed whether sign-up is fully open for every account or region, so check the console directly.
[VERIFY 2026-06-22.]
Can I use Fugu from the EU?
There are reports of an EU and EEA availability restriction at launch, but this has not been confirmed against an official statement. If you are in those regions and the console blocks sign-up or billing, that restriction is the likely reason. Verify Sakana’s terms and the console behavior before planning a rollout.
Do I need a special SDK to call Fugu?
No. Fugu exposes an OpenAI-compatible endpoint, so you can point an existing OpenAI client at it with your API key and the base URL from your console. No SDK migration is required. See how to use the Sakana Fugu API for the full setup.
What is the difference between Fugu and Fugu Ultra at sign-up?
Both variants live behind the same login and endpoint. You do not buy them separately at the account level. You choose between them per request by setting the model ID.
Use fugu for balanced, lower-latency workloads. Use fugu-ultra when answer quality matters more than speed or cost. The beta name “Fugu Mini” maps to today’s “Fugu.”
Bottom line
To access Sakana Fugu, sign in at console.sakana.ai, verify whether your account can create keys, copy the console-provided base URL, and call the OpenAI-compatible endpoint.
The cost answer is also clear: there is no standalone free tier. Verify the paid tiers and any launch promo directly in the console before budgeting.
Once your key is live, set up your first Fugu request in Apidog so testing, sharing, and documentation start before production integration.



Top comments (0)