Sao10k API

Llama 3.1 70B Hanami x1 is an experimental 70B model built on top of Euryale v2.2 by Sao10K, offering a different feel with enhanced creativity and logical reasoning. The creator considers it an improvement over both Euryale v2.1 and v2.2.

Llama 3.3 Euryale 70B v2.3 is the latest in Sao10K's Euryale series, built on Llama 3.3 Instruct with a 131K context window and 16K output limit. It's a direct successor to v2.2, trained without LoRA extraction for more robust creative roleplay and storywriting performance.

Llama 3.3 Euryale 70B v2.3 is a creative writing and roleplay finetune of Meta's Llama 3.3 70B Instruct by Sao10K, an independent model creator whose Euryale line is their flagship series for these tasks. v2.3 is the successor to Euryale v2.2. Unlike earlier versions, it is trained entirely on top of Llama 3.3 Instruct rather than LoRA-extracted, and the creator describes it as a direct replacement with improved performance. Its training data mixes roleplay and creative content with instruction data, so it handles storytelling, interactive dialogue, and general instruction following. The model uses the Llama 3 Instruct prompt format, and the creator recommends a temperature around 1.1 with min_p 0.1. Through the API it offers a 131K context window with up to 16.4K output tokens, a fit for long multi-turn roleplay and story sessions.

72B Qwen2.5 Kunou v1 is a roleplay and creative writing model by Sao10K, the independent creator behind the Euryale and Stheno series, finetuned from Qwen2.5 72B Instruct. Sao10K describes it as a generalist focused on roleplay and creative instruction tasks, and as something of a successor to Llama 3 Euryale v2.2, trained on a cleaned-up version of the datasets used for Euryale and Stheno. The model uses the ChatML prompt format. The creator notes that system prompts have a strong effect on its behavior, and recommends a temperature of 1.1 with min_p 0.1. It is the largest model in the Kunou series (14B and 32B versions also exist) and suits developers who want Euryale-style roleplay and storytelling on a Qwen base instead of Llama.

Llama 3.1 Euryale 70B v2.2 is Sao10K's creative roleplay model built on Meta's Llama 3.1 architecture with improved multi-turn coherency, system prompt handling, and reasoning capabilities. It features a 32K context window and excels at immersive storytelling with strong prompt adherence.

Llama 3.1 Euryale 70B v2.2 is a creative writing and roleplay fine-tune of Meta's Llama 3.1 70B by Sao10K, an independent model creator known for the Euryale and Stheno series. The model was trained in two epochs, first on multi-turn conversational instruction data, then on creative writing and roleplay content mixed with creative instruction sets. Compared to v2.1, Sao10K added 55% more roleplaying examples and 40% more creative writing examples, replaced single-turn prompts with responses from Claude 3.5 Sonnet and Claude 3 Opus, and added datasets targeting system prompt adherence and reasoning. It uses the Llama 3 Instruct prompt format, and the creator recommends a temperature around 1.2 with min_p 0.2. Through the API it offers a 131K context window with up to 16.4K output tokens, suited to long multi-turn roleplay and story sessions.

Llama 3 8B Lunaris is a versatile 8B parameter generalist and roleplaying model created by merging five different Llama 3-based models. It balances creativity with improved logical reasoning and general knowledge, serving as an evolution of Stheno v3.2.

Llama 3 Euryale 70B v2.1 is a 70-billion parameter model by Sao10K focused on creative roleplay and storytelling, featuring strong prompt adherence, spatial awareness, and non-restrictive creative writing capabilities. It adapts well to custom formatting and produces highly varied, creative outputs.

Llama 3 Euryale 70B v2.1 is a creative roleplay and storywriting model from Sao10K, an independent model creator, built as a fine-tune of Meta's Llama 3 70B. It is the larger sister model to Sao10K's popular Stheno 8B, trained with LoRA on the same dataset as Stheno v3.2. Compared to earlier Euryale releases, v2.1 improves prompt adherence and spatial awareness, adapts better to custom reply formats, and produces varied, creative outputs without being restrictive during roleplay. The creator recommends sampler settings of temperature 1.17 and min_p 0.075. The model has an 8K context window and an 8K output limit, which suits shorter scenes and chats. For longer context, Sao10K's later Llama 3.1 Euryale v2.2 and Llama 3.3 Euryale v2.3 offer 131K.