ReadyArt/Serenity-27B

VISIONPricing:Input $1.6 / Cached $0.15 / Output $12Concurrent Unit Cost:2Model Size:27BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Aug 18, 2026License:apache-2.0Architecture:Transformer0.0K Open Weights Featherless Exclusive Cold

ReadyArt/Serenity-27B is a 27 billion parameter LoRA fine-tuned model based on Qwen3.8, developed by ReadyArt and contributors like FrenzyBiscuit. It merges the expressive depth of Melody and immersive presence of Darkside lineages, optimized for adult-oriented, uncensored roleplay with a 32768 token context length. This model is specifically designed for personal, private use in chat completion mode with strict post-processing and uncensored prompts.

Loading preview...

Overview

Serenity-27B is a 27 billion parameter model from ReadyArt, built upon the Qwen3.8 architecture and fine-tuned using LoRA (Low-Rank Adaptation) at an unprecedented rank of 160. This model uniquely harmonizes two distinct lineages: Melody (male → female perspective) and Darkside (female → male perspective), aiming for a balanced and deeply immersive roleplay experience. It is specifically designed for adult-oriented, uncensored interactions, leveraging advanced "Character Engine" and "Emotional Engine" datasets to ensure consistent personality, dynamic emotional states, and high-quality, non-repetitive dialogue.

Key Capabilities

  • Dual-Lineage Integration: Combines male → female and female → male perspectives for a rich, balanced tone.
  • Immersive Roleplay: Optimized for mature themes and uncensored dialogue, creating genuine life and emotional resonance.
  • Consistent Characterization: "Character Engine" ensures consistent personality traits and speech patterns.
  • Dynamic Emotional Responses: "Emotional Engine" injects realistic emotional states into interactions.
  • High-Quality Output: Automated refinement processes remove repetitive phrases and maintain dialogue integrity.
  • Extended Context: Supports a 32768 token context length for longer, more complex interactions.

Important Usage Requirements

  • Chat Completion Mode: Must be used with chat completion APIs (e.g., /v1/chat/completions), not text completion.
  • Strict Post-Processing: Requires specific prompt post-processing settings (e.g., Strict (user first, alternating roles; with tools) in SillyTavern).
  • Uncensored Prompts: Demands fully uncensored system prompts and character cards; safety-aligned prompts will degrade performance.
  • Personal Use Only: Strictly for personal, private use; not intended for public API deployment due to its unaligned and uncensored nature.

Finetunes
5 models