No Claude Fable 5? No problem: Sakana achieves frontier performance with new Fugu multi-model, auto synthesis system

2026-06-22 · Source: VentureBeat · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering, Cloud Computing & IT Infrastructure · Depth: Intermediate, long

Summary

Sakana, an AI startup, launched Fugu on June 22, 2026, a multi-agent orchestration system providing frontier-level AI performance through an OpenAI-compatible API. Designed to counter vendor lock-in and export controls, Fugu dynamically routes queries to a swappable pool of specialized AI agents, bypassing traditional monolithic models. This system, based on Sakana's TRINITY and Conductor research, functions as a master contractor, breaking down complex requests, delegating sub-tasks, and synthesizing final outputs. Fugu offers two variants: Fugu for high-speed, low-latency tasks and Fugu Ultra for complex, high-stakes applications like AI research. Fugu Ultra achieved 93.2 on LiveCodeBench and 95.5 on GPQA-D, surpassing Anthropic's Fable 5 and Mythos Preview, and scored 73.7 on SWE-Bench Pro, beating Claude Opus 4.8 (69.2) and GPT-5.5 (58.6). It is a proprietary API service with specific pricing, including \$5 per million input tokens and \$30 per million output tokens for Fugu Ultra, and is currently restricted in the EU/EEA.

Key takeaway

For AI Engineers and MLOps teams building critical enterprise workflows, Sakana's Fugu offers a compelling hedge against single-vendor AI reliance and geopolitical risks. You should evaluate Fugu's multi-agent orchestration for complex, multi-step tasks where resilience and dynamic model access are paramount. Its ability to match or exceed frontier model performance on benchmarks like LiveCodeBench and SWE-Bench Pro, while abstracting orchestration complexity, makes it a strong candidate for your next-generation AI infrastructure, despite its premium pricing and EU/EEA restrictions.

Key insights

Multi-agent orchestration provides resilience and competitive performance by dynamically routing tasks to specialized AI agents.

Principles

• Collective intelligence hedges against single-vendor reliance.
• Orchestration models represent the next frontier.
• Dynamic agent pools route around vendor restrictions.

Method

The system breaks problems into sub-tasks, delegates to expert foundation models, verifies work, and synthesizes output using learned coordination strategies.

In practice

• Apply Fugu for complex, multi-step tasks requiring delegation.
• Configure routing pools to meet corporate privacy standards.
• Utilize Fugu Ultra for high-stakes AI research or cybersecurity analysis.

Topics

• Sakana Fugu
• Multi-agent Orchestration
• AI Infrastructure Resilience
• Vendor Lock-in
• LLM Benchmarking
• API Services

Best for: CTO, VP of Engineering/Data, Machine Learning Engineer, AI Engineer, MLOps Engineer, Director of AI/ML

Related on AIssential

• Sakana Fugu: A Multi-Agent Orchestration System as a Foundation Model — Multi-agent orchestration dynamically coordinates foundation models for superior performance in complex tasks.
• ChinAI #355: An Alliance for AI's "Harness Era" -MiniMax + Alibaba Cloud — The "Harness Era" of AI agents demands new cloud infrastructure to overcome critical engineering barriers for enterprise adoption.
• NVIDIA GTC 2026 Open Models Panel Highlights with Jensen Huang — AI's future lies in a "proprietary and open" ecosystem, leveraging compound agents and orchestrated multi-model systems.
• SUPERLINEAR RAISES €6M TO BUILD THE AI OPERATING SYSTEM FOR ENTERPRISE ORCHESTRATION — Superlinear's Holon AI OS orchestrates enterprise operations for 10-30% productivity gains, aiming to boost EU GDP by 1%.
• Announcing a Strategic Partnership with Datadog — Sakana AI and Datadog announced a strategic partnership on February 26, 2026, combining Sakana AI's research in efficient, scalable, and adaptive AI models with Datadog's platform for observability and security.
• Sakana AI's Fugu orchestrates multiple LLMs to match Anthropic's Fable and Mythos benchmarks — Fugu orchestrates multiple LLMs to match top benchmarks and reduce single-provider dependence.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by VentureBeat.