2 hours ago
1
Sakana AI just dropped what might be the most interesting architectural bet in the AI race right now. The Tokyo-based research lab unveiled Sakana Fugu, a multi-agent orchestration system that doesn’t try to be the biggest model in the room. Instead, it conducts a symphony of specialist models, including third-party frontier LLMs, through a single OpenAI-compatible API.
The headline number: Fugu Ultra, the system’s top-tier variant, scored 73.7 on the SWE-Bench Pro benchmark. That reportedly matches or approaches the performance of Anthropic’s Fable 5 and Mythos Preview. In English: a system that coordinates multiple AI agents is now performing on par with some of the most powerful standalone models on the planet.
How Fugu actually works
Fugu operates as a trained conductor model that manages an entire orchestra of specialist AI agents. Unlike traditional multi-agent setups that rely on fixed, hard-coded workflows, Fugu dynamically handles routing, role assignments, verification steps, and result synthesis. The system learns collaboration patterns autonomously, deciding in real-time which specialist model should handle which part of a complex task.
The whole thing is delivered through a single API endpoint that’s compatible with OpenAI’s format. For developers, that means integration doesn’t require rearchitecting existing applications. You swap out an endpoint and suddenly have access to a coordinated multi-agent system instead of a single model.
The strategic calculus behind the architecture
Fugu isn’t just a technical achievement. It’s a strategic positioning play that addresses two of the most pressing concerns in enterprise AI adoption: export control risk and vendor lock-in. By orchestrating multiple models from different providers, Sakana AI is essentially building redundancy into the AI stack. If one model provider faces regulatory restrictions, gets acquired, or simply raises prices, the system can route around the problem.
Fugu is available globally with one notable exception. The EU and EEA are temporarily locked out pending regulatory compliance.
The development of Fugu builds on Sakana AI’s earlier research through the TRINITY and Conductor projects, both of which explored learned orchestration concepts.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.
English (US) · 