Nabbers1999/Bathory-X-26B-A4B-0.2

Hugging Face
VISIONPricing:Input $0.4 / Cached $0.08 / Output $0.8Concurrent Unit Cost:2Model Size:26BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Oct 12, 2026License:apache-2.0Architecture:Transformer0.0K Open Weights Featherless Exclusive Warm

Nabbers1999/Bathory-X-26B-A4B-0.2 is a 26 billion parameter experimental language model with a 32768 token context length, developed by Nabbers1999. This model is a result of "model surgery," combining layers from multiple existing models and distilling its language model head from another via top-k logit training. It is specifically optimized for improved writing skills, aiming to avoid common linguistic cliches and repetitive phrasing, making it suitable for creative text generation and roleplay scenarios.

Loading preview...

Bathory X 26B-A4B v0.2 Experimental Model

Bathory X 26B-A4B v0.2 is an experimental 26 billion parameter language model developed by Nabbers1999, representing a significant improvement over its predecessor. This model is a unique "Frankenstein" creation, utilizing a technique called model surgery, which involves merging and directly editing layers from several different base models.

Key Capabilities

  • Enhanced Writing Skills: Specifically designed to produce more solid writing, avoiding common cliches, similes, and repetitive phrasing.
  • Model Surgery Approach: Built through a novel process of combining body layers from models like WaifuGemma4-26b-a4b-v1, Orion-26B-A4B-v1.1, For-Her-Darkside-26B-A4B-v1.4, and Dark-Scarlett-v1.0-26B-A4B.
  • Distilled LM Head: The language model head was distilled using top-k logit training from Gryphe/Gemma-4-26B-A4B-StyleTune-V2, focusing on prose and roleplay datasets.
  • Task Arithmetic Integration: Edits from coder3101/gemma-4-26B-A4B-it-heretic were combined via task arithmetic to further refine its capabilities.

Good For

  • Creative Text Generation: Ideal for applications requiring original and nuanced prose.
  • Roleplay Scenarios: Excels in generating engaging and less formulaic dialogue and narratives for roleplaying.
  • Exploratory AI Research: Useful for researchers interested in model merging and surgical techniques for LLM development.