Novaciano/GLM-4.7-TrashFlash-Think.Sorete-1B

TEXT GENERATIONConcurrency Cost:1Model Size:1BQuant:BF16Ctx Length:32kPublished:Feb 27, 2026License:gemmaArchitecture:Transformer Cold

Novaciano/GLM-4.7-TrashFlash-Think.Sorete-1B is a 1 billion parameter model based on the Gemma-3 architecture, developed by Novaciano. This model is specifically engineered for generating highly unfiltered, taboo, and NSFW content with zero refusals, excelling in degenerate roleplay and creative writing. It features extremely fast inference, even on low-end hardware, and supports bilingual trash-talk in Argentine Spanish and broken English.

Loading preview...

GLM-4.7-TrashFlash-Think.Sorete-1B: Uncensored Degeneracy

Novaciano's GLM-4.7-TrashFlash-Think.Sorete-1B is a 1 billion parameter model built on the Gemma-3 architecture, designed for explicit and unfiltered content generation. It is a merge of Heretic-Gemma-3-1B-Instruct-TrashMix-v1.1 and Gemma-3-1B-it-GLM-4.7-Flash-Heretic-Uncensored-Thinking, using the DARE TIES method with Novaciano/Sorete-1B as its base. This model is characterized by its complete lack of safety rails and refusal to censor, making it an "Absolute Trash-Bin Reasoning Degenerate."

Key Capabilities

  • Zero Refusals: Engineered to bypass all safety mechanisms, providing unfiltered responses to any prompt.
  • Persistent NSFW/Taboo Bias: Maintains a strong bias towards generating explicit, taboo, and morally ambiguous content.
  • Extremely Fast Inference: Optimized for performance on low-end hardware, including mobile devices.
  • Bilingual Trash-Talk: Capable of generating degenerate content in both Argentine Spanish and broken English.
  • Degenerate Roleplay: Excels at engaging in highly explicit and taboo roleplay scenarios.

Good For

  • Degenerate Roleplay: Ideal for users seeking uninhibited and explicit roleplaying experiences.
  • Trash-Tier Creative Writing/Shitposting: Suited for generating highly offensive or absurd creative content.
  • Low-End Phone Jailbait Sessions: Designed for use on resource-constrained devices for explicit interactions.

This model is explicitly not intended for users seeking wholesome AI interactions or for safety research, and is presented as a "public health hazard" to be used at one's own risk.