Sakana: Fugu Ultra
Fugu Ultra is the higher-performance model in Sakana AI's Fugu family. Rather than a single monolithic model, Fugu is a learned multi-agent orchestration system: a language model trained to route...
Anyone in the Space can @-mention Sakana: Fugu Ultra with the team's shared context - pooled credits, one chat, one memory.
Starter is free forever - 1 Space, 100 credits/month, 1 MCP. No card.
Verdict
Best for
- Processing entire codebases in one context
- Multi-document analysis and synthesis
- Long-form content generation with deep context
- Screenshot and document vision tasks
- Research workflows requiring extensive reference material
Strengths
The 1M token context window handles workloads that would require chunking or RAG pipelines in smaller models — paste an entire repository or a stack of PDFs and work directly with the full content. Vision support extends this to screenshots and scanned documents, useful for UI analysis or extracting data from images. The $5 input pricing makes exploratory passes through large documents economical compared to premium-tier alternatives.
Trade-offs
No public benchmark data means you're flying blind on reasoning quality, code generation accuracy, and how it stacks up against Claude or GPT-4 class models on standard evals. The $30 output cost per Mtok adds up fast if you're generating long responses — a 10k token summary costs $0.30, triple what you'd pay with some competitors. Sakana is a newer player without the ecosystem maturity of Anthropic or OpenAI, so expect less community tooling and fewer integration examples.
Specifications
- Provider
- sakana
- Category
- llm
- Context length
- 1,000,000 tokens
- Max output
- 128,000 tokens
- Modalities
- text, image
- License
- proprietary
- Released
- 2026-06-24
Pricing
- Input
- $5.00/Mtok
- Output
- $30.00/Mtok
- Model ID
sakana/fugu-ultra
Per-token prices show what the model costs upstream. On Switchy your team draws from one shared org credit pool - one plan, one balance for everyone.
Team cost calculator
5 seats · 80 msgs/day
Switchy meters this against your org's shared credit pool - one plan, one balance for everyone.
Providers
| Provider | Context | Input | Output | P50 latency | Throughput | 30d uptime |
|---|---|---|---|---|---|---|
| sakana | 1000k | $5.00/Mtok | $30.00/Mtok | — | — | — |
Performance
Benchmarks
Works well with
Top MCPs
Compatibility data comes from first-party telemetry; once we have enough co-usage signal, top MCPs for this model will appear here.
How Switchy teams use it
Starter prompts
Codebase Architecture Review
Review this codebase and provide a high-level architecture summary. Identify the main modules, their responsibilities, and any patterns or anti-patterns you observe. Focus on how components interact and where technical debt might be accumulating.Open in a Space →
Multi-Document Synthesis
I've provided five research papers on the same topic. Identify the key themes across all papers, note where findings agree or conflict, and summarize the current state of research in this area.Open in a Space →
Screenshot UI Analysis
Analyze these application screenshots and document the UI patterns used. Identify consistency issues, accessibility concerns, and suggest specific improvements to the design system.Open in a Space →
Contract Comparison
Compare these three contracts and create a table showing where terms differ. Highlight any clauses that appear in one version but not others, and flag terms that might create conflicts.Open in a Space →
Long-Form Content Outlining
Using the research materials provided, create a detailed outline for a comprehensive guide on this topic. Include chapter summaries, key points for each section, and notes on where specific sources support each claim.Open in a Space →