Sakana Chat rolls out Fugu Max model and memory feature

Sakana AI's consumer chat product now runs on Fugu Max, and it will remember what you tell it. In a product post dated September 17, 2026, the Tokyo lab said Sakana Chat had rolled the Fugu Max orchestrator model out to all users, alongside a new cross-conversation memory feature.

Fugu Max is billed as an "orchestrator model" that "brings together a diverse pool of open models and routes each prompt to the model best suited to it." The pitch is frontier-adjacent quality without tying the product to any one lab's model.

Memory, in the company's words, means "what you tell it once (your role, the project you are working on, how you like answers formatted and written) carries over into later conversations." The setting applies only to chats after this update; users can inspect the stored items in settings or disable the feature entirely. Both Namazu and Fugu Max support it.

Sakana Chat launched with the Namazu alpha in March 2026, then added Sakana Fugu, code execution and file attachments in August. The update names no benchmarks and lists no pricing for the chat itself.