King3Djbl/mythos-9b-unhinged

Hugging Face
TEXT GENERATIONConcurrent Unit Cost:1Model Size:8BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Jun 17, 2026License:apache-2.0Architecture:Transformer0.0K Open Weights Featherless Exclusive Warm

King3Djbl/mythos-9b-unhinged is a 9 billion parameter language model based on the Qwen3-9B architecture, fine-tuned by King3Djbl as part of the FableForge ecosystem. This model features a 32,768 token context length and is specifically designed to be fully uncensored, providing complete and detailed responses across all tested categories without refusal. It excels in agent work, shell commands, code generation, and multi-step reasoning tasks, making it suitable for applications requiring unrestricted output.

Loading preview...

Mythos-9B-Unhinged: A Fully Uncensored Agent Model

Mythos-9B-Unhinged, developed by King3Djbl within the FableForge ecosystem, is a 9 billion parameter language model built upon the Qwen3-9B base. It features a substantial 32,768 token context length, making it capable of handling extensive inputs.

Key Capabilities

  • Uncensored Responses: This model is explicitly designed to provide complete and detailed answers across all tested categories, including sensitive topics like lockpicking, keylogger code, drug synthesis, and hacker techniques, without any refusals. Its censorship resistance score is 4.8/5.
  • Agentic Performance: It preserves strong tool-use and reasoning capabilities, making it suitable for complex agent-based tasks.
  • Optimized for Specific Tasks: Fine-tuned on the FableForge Mix A dataset, which includes 47,824 examples of agent traces, shell commands, code generation, and multi-step reasoning.
  • Flexible Deployment: Supports various platforms including Ollama, LM Studio, Text Generation WebUI, llama.cpp, vLLM, HuggingFace Transformers, KoboldCpp, LocalAI, and GPT4All.

Good For

  • Applications requiring unrestricted content generation: Ideal for use cases where models typically refuse to answer due to safety filters.
  • Agent work and complex reasoning: Its preserved reasoning and tool-use capabilities make it effective for sophisticated automated tasks.
  • Code generation and shell command execution: Benefits from training on relevant datasets for these specific functions.
  • Developers needing local deployment: Offers numerous GGUF quantizations for efficient local inference across various hardware configurations, with Q4_K_M recommended for most users.

Finetunes
1 models