by Sao10kReleased 8 January 2025
This is [Sao10K](/sao10k)'s experiment over [Euryale v2.2](/sao10k/l3.1-euryale-70b).
Reads about 16 pages
Standard OpenAI fields, grouped by what they control. Anything not listed is accepted and ignored.
Shaping the reply
How long, how varied, where to stop.
max_tokenstemperaturetop_ptop_kmin_pstopseedfrequency_penaltypresence_penaltyrepetition_penaltylogit_biasStructured output
Get JSON you can parse instead of prose.
structured_outputsNo. Sao10k has retired it. Other Sao10k models are on sale.
The lab has not said.
Sao10k names no languages for Llama 3.1 70B Hanami x1. Arabic, French and Darja: not stated.
It reads text. Up to 16K tokens in one request, about 16 pages. Longest reply: not stated.
Yes, its weights are public. Open weights on Hugging Face
Sao10k. Released 8 January 2025. It knows the world up to 31 December 2023.
Lunaris 8B is a small, general-purpose and roleplay model built on Llama 3, blended from several models to balance creative writing with sound logic and general knowledge. It reads and writes text and keeps about 8 pages of conversation in mind. Its maker presents it as an upgrade on Stheno v3.2 and suggests a high creativity setting of 1.4.
Sao10k's Euryale L3.1 70B v2.2, a creative roleplay model on Llama 3.1 that follows v2.1, with tool calling and a 128K context; Euryale L3.3 is the later version
From what its maker and two benchmark sites publish, and where it stands on OpenDunes. Read 7 Nov 2026.
Thinks first No.
Works through the problem before it answers.
Uses your tools No.
Can call functions your code provides.
Fills your JSON
Answers in the JSON shape you give it.
Searches the web No.
Can look things up online while it answers.
Cheaper on repeats No.
A prompt it saw a moment ago costs less the second time.
Open weights
The model's weights are public.
Knows events up to 31 Dec 2023
What it learned stops at this date.