Fugu Ultra
Fugu Ultra is a large language model made by Sakana, added to the catalog on 24 Jun 2026. It has a context window of 1,000,000 tokens and costs $5.00 per 1M input tokens and $30.00 per 1M output tokens. It accepts text and images and produces text. Typical uses: working with long text and image prompts, image-and-text question answering and structured text generation.
Price checked ยท source: OpenRouter
What Fugu Ultra is
Sakana: Fugu Ultra is a text-generating model from Sakana. Its catalog description presents Fugu as a learned multi-agent orchestration system rather than a single monolithic model, and identifies Fugu Ultra as a model in the Fugu family. The catalog places the family within Sakana AI's offerings. The available description does not specify the routing process in further detail. The model accepts text and images and produces text, so its interaction can combine written prompts with visual input while returning a text response.
It has a context window of up to 1,000,000 tokens and a maximum output of 128,000 tokens. This gives it room to receive extensive text or mixed text-and-image material in one request, and to return long text outputs within the stated limit. The catalog lists support for tools, structured outputs, reasoning, and web search options. These are supported capabilities; the catalog does not name particular tools, define a structured-output schema, or give details about the web-search setup. Fugu Ultra may suit work that combines long-context reading, image inputs, and text generation, as well as workflows that need tool support or structured responses. The listed features describe capabilities, not a guarantee of any particular result. It produces text only; image is an input modality, not a listed output format.
Key features
- Accepts text and image inputs, and produces text.
- Supports a context window of up to 1,000,000 tokens.
- Maximum output length is 128,000 tokens.
- Supports tools, structured outputs, reasoning, and web search options.
Typical uses
- Working with long text and image prompts
- Image-and-text question answering
- Structured text generation
- Tool-supported research
- Reasoning over lengthy inputs
Fugu Ultra pricing
US dollars, list price through OpenRouter.
| Item | Standard |
|---|---|
| Input, per 1M tokens | $5.00 |
| Output, per 1M tokens | $30.00 |
| Cached input read, per 1M tokens | $0.50 |
| Web search, per search | $0.01 |
| Input above 272,000 prompt tokens, per 1M | $10.00 |
| Output above 272,000 prompt tokens, per 1M | $45.00 |
No price change since it was added on 24 Jun 2026.
| What a task costs | Cost |
|---|---|
| Answer a short question500 in / 300 out | $0.012 |
| Summarise a 50-page PDF25,000 in / 800 out | $0.149 |
| Review a 2,000-line source file30,000 in / 2,000 out | $0.210 |
| Read a 300-page contract set150,000 in / 1,500 out | $0.795 |
Fugu Ultra specifications
| Model id | sakana/fugu-ultra |
|---|---|
| Version | fugu-ultra-20260615 |
| Context window | 1,000,000 tokensAbout 750,000 English words |
| Max output | 128,000 tokens |
| Knowledge cutoff | Not published |
| Added to catalog | 24 Jun 2026 |
| Input types | text, image |
|---|---|
| Output types | text |
| Tool calling | Yes |
| Structured outputs | Yes |
| Reasoning effort control | Yes |
| Content moderation | Not applied by provider |
Sample outputs
The same prompts are given to every model, and the unedited output is shown with the date it was produced. Upend does not publish leaderboard scores it did not run.
No output has been recorded for Fugu Ultra yet.
Compare Fugu Ultra
Side by side with other models: prices, context window, what each accepts and supports, and the same four tasks priced on both.