Sakana AI has unveiled Fugu Ultra, an orchestration layer that assembles and routes subtasks across a pool of models th…

Rohan Paul Twitter · Rohan Paul (@rohanpaul_ai) · 2026-06-22

Sakana AI launches Fugu Ultra, an orchestration layer that uses a learned coordinator model to route subtasks across a pool of AI models through a single OpenAI-compatible endpoint, matching benchmark performance of competing systems Fable and Mythos.

Open original ↗

Appears in

Sakana AI Fugu Ultra: Multi-Model Orchestration Layer Launch and Early Benchmarks

Extraction

Topics: ai-orchestrationmulti-agent-systemsmodel-routingllm-infrastructure

Claims

Fugu Ultra is an orchestration layer that decomposes and routes subtasks across a pool of models via a single OpenAI-compatible endpoint.
Fugu contains a learned coordinator model that manages routing decisions rather than relying on static rules.
Fugu Ultra matches the benchmark performance of competing systems Fable and Mythos on most evaluations.

Key quotes

Sakana AI has unveiled Fugu Ultra, an orchestration layer that assembles and routes subtasks across a pool of models through one OpenAI-compatible endpoint.

Fugu is a learned coordinator model inside a [system].