LLM Reference

Fugu Ultra

Released
2026-06-22
Last refreshed
2026-06-22
Status
Researched 1d ago
ProprietaryCommercial use: conditionalMultimodalCodingRAGAgentsLong contextVisionJSON / Tool use

Fugu Ultra is worth evaluating for coding, rag, and agents when its provider route and context window match the workload.

Use it for

  • Teams evaluating coding, rag, and agents
  • Workloads that can use a 1m context window
  • Buyers comparing 2 tracked provider routes

Do not use it for

  • Workloads where another current model has stronger sourced task evidence
Specifications
Family
Fugu
Released
2026-06-22
Context
1m
Architecture
Composition of Experts
Specialization
general
Openness
Proprietary
License
ProprietaryCommercial use: conditional
Training
Reinforcement Learning
Created by

Nature-inspired AI foundation models.

Tokyo, Japan
Founded 2023
Website
Pricing
Output / 1M
$30.00
Input / 1M
$5.00

Cheapest of 2 routes · Sakana AI · cache read $0.500

About

Fugu Ultra is Sakana AI's quality-optimized multi-agent orchestration model. Achieves SWE-Bench Pro 73.7, LiveCodeBench 93.2, GPQA-D 95.5. Supports extended context beyond 272K tokens at higher pricing. Versioned API ID: fugu-ultra-20260615. Supports text and image input, reasoning effort levels (high, xhigh/max), structured outputs, and function calling via OpenAI-compatible API (base_url: https://api.sakana.ai/v1). Not available in EU/EEA.

Fugu Ultra is a proprietary model in the Fugu family. The structured metadata tracks a 1m-token context window, multimodal input, reasoning, function calling, tool use, and structured outputs. This page tracks provider routes through Sakana AI and Vercel AI Gateway, with the cheapest tracked route listed at $5 input and $30 output per 1M tokens. Headline tracked benchmarks include SWE-bench Pro 73.7, Terminal-Bench 2.1 82.1, and LiveCodeBench 93.2.

Top use-case fit: coding, agents, and build tasks

Coding

Q/$ D

2 relevant benchmarks in the decision map.

RAG

Included by capability and metadata signals in the decision map.

Agents

Included by capability and metadata signals in the decision map.

Provider price ladder

Compare all 2

Compare API pricing across 2 providers for input and output tokens, batch, and cached reads when available.

ProviderInput / 1MOutput / 1MCacheRoute
Sakana AI$5.00$30.00read $0.500
Serverless
Vercel AI Gateway$5.00$30.00read $0.500
Serverless

Capabilities

VisionMultimodalReasoningFunction CallingTool UseStructured Outputs

Benchmark peer barsfor Coding

Benchmark scores(6)

Scores are benchmark-specific and are direction-aware: the same numeric gap can mean very different outcomes across suites. Use the leaderboard context and this model's provider route to decide whether the winning margin is meaningful for your workload.
BenchmarkScoreVersionSource
SWE-bench Pro73.7SWE-bench Pro, Sakana AI-reportedhttps://sakana.ai/fugu/
Terminal-Bench 2.182.1Terminal-Bench 2.1, Sakana AI-reportedhttps://sakana.ai/fugu/
LiveCodeBench93.2LiveCodeBench, Sakana AI-reportedhttps://sakana.ai/fugu/
Google-Proof Q&A95.5diamondhttps://sakana.ai/fugu/
Humanity's Last Exam50.0Humanity's Last Exam, Sakana AI-reportedhttps://sakana.ai/fugu/
CharXiv86.6reasoninghttps://sakana.ai/fugu/

Migration checks

No linked migration route is available for this model yet.