According to emollick, Fugu Ultra-high took 30 minutes on shader tests and trailed Fable quality, despite Sakana claiming Fable level performance.

According to KyeGomezB, Sakana AI’s Fugu Ultra routes subtasks across multiple LLMs via an OpenAI API endpoint, matching Fable and Mythos on benchmarks.

According to TheRundownAI, Sakana AI’s Fugu Ultra uses multi agent orchestration to match Fable and Mythos on several benchmarks.