Sao10K: Llama 3.3 Euryale 70B

Sao10K: Llama 3.3 Euryale 70B is a community-built fine-tune in the Euryale series, positioned as the direct successor to Euryale L3 70B v2.2 and designed primarily around creative roleplay. Rather than a general-purpose assistant, it targets narrative generation, character voice, and other story-driven workloads where sustained persona quality and long conversational memory matter more than raw factual recall. Building on the Llama 3.3 70B base gives the model room to keep extended scenes coherent while remaining lightweight enough to run on a single high-memory node.

Practically, the model fits use cases such as interactive fiction, character-driven chat, and longer collaborative writing sessions where roleplay consistency over many turns is the main quality bar. The 131,072-token context window allows very long shared histories to stay in scope, which is especially useful for novel-length or campaign-style interactions that would lose detail in shorter-context alternatives. For teams already invested in open Llama-family deployments who want a roleplay-specialized upgrade without leaving that ecosystem, Euryale L3.3 70B offers a focused, drop-in option.