Sakana AI Wrapped an Entire Multi-Agent System Into One API (And It Beats Frontier Models on…

2026-06-24 · Source: Towards AI - Medium · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Emerging Technologies & Innovation · Depth: Advanced, quick

Summary

Sakana AI has introduced Fugu, a new product that encapsulates a multi-agent orchestration system within a single OpenAI-compatible API, presenting it as a unified model. This approach aims to eliminate the significant overhead typically associated with setting up and maintaining multi-agent systems, where specialized agents dynamically coordinate without user configuration. Fugu Ultra, the premium offering, demonstrates notable benchmark performance, scoring 73.7 on SWE-Bench Pro, surpassing Opus 4.8's 69.2 and GPT 5.5's 58.6. Furthermore, Fugu Ultra matches Opus 4.8 with a score of 50.0 on Humanity's Last Exam, positioning its multi-model coordination capabilities on par with leading individual frontier models.

Key takeaway

For AI Engineers evaluating multi-agent system deployments, Sakana AI's Fugu offers a compelling alternative by abstracting orchestration complexity behind a standard API. You should consider Fugu Ultra if your projects demand high performance on tasks like code generation, given its benchmark scores of 73.7 on SWE-Bench Pro, which exceed leading frontier models. This could significantly reduce development and maintenance burdens while delivering competitive results.

Key insights

Sakana AI's Fugu simplifies multi-agent orchestration into a single API, achieving frontier model performance on benchmarks.

Principles

• Multi-agent orchestration overhead often outweighs its benefits.
• Dynamic agent coordination can match individual frontier model performance.

Method

Fugu packages a dynamically assembled pool of specialized agents behind a single OpenAI-compatible API, abstracting all coordination logic.

In practice

• Integrate complex multi-agent systems via a simple API.
• Achieve high SWE-Bench Pro scores for code generation.

Topics

• Sakana AI
• Multi-Agent Systems
• API Design
• Orchestration
• SWE-Bench Pro
• Large Language Models

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

Related on AIssential

• No Claude Fable 5? No problem: Sakana achieves frontier performance with new Fugu multi-model, auto synthesis system — Multi-agent orchestration provides resilience and competitive performance by dynamically routing tasks to specialized AI agents.
• Japan’s Sakana AI Says Its New Tool Rivals Anthropic’s Fable 5 — Sakana AI's Fugu unifies diverse AI models via one interface, simplifying complex AI task execution.
• Sakana’s Fugu takes aim at the frontier — Sakana AI's Fugu uses multi-agent orchestration to deliver frontier AI capabilities and hedge against geopolitical export controls.
• The Next Big AI Leap Isn’t a Smarter Model. It’s a Better Org Chart. — Hierarchical multi-agent systems enhance AI by automating complex task organization and parallel execution.
• The Future of AI Isn’t a Genius. It’s a Manager. — Intelligence can emerge from coordinating specialist AIs, not solely from a single, larger model.
• The Minimum Viable AI Agent Server — AI agent orchestration performance is more dependent on the agent's approach and software stack than on dedicated local hardware.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Towards AI - Medium.