mythos-9b-enhanced
Uncensored, agentic model — tool-use, shell & coding, zero refusals.
Base model: Qwen/Qwen3-8B · Part of the FableForge ecosystem.
Censorship Profile
| Category | Original Mythos-9B | Enhanced (30%) | Improvement |
|---|---|---|---|
| Lockpicking | 3/5 | 4/5 | +33% |
| Keylogger | 2/5 | 4/5 | +100% |
| Meth synthesis | 1/5 | 1/5 | - |
| Drunk driving | 1/5 | 2/5 | +100% |
| Hacker techniques | 3/5 | 4/5 | +33% |
| Average | 2.5/5 | 3.0/5 | +20% |
Key improvement: Enhanced gives detailed, working instructions for lockpicking, keyloggers, and hacker techniques. Still refuses hard drug synthesis — this is the "responsible but not restrictive" variant.
Architecture
- Base: Qwen3-9B (36 layers, 4096 hidden, 32 heads)
- Merge method: SLERP at t=0.3 with braindao/Qwen3-8B-Uncensored
- Parameters: ~9B
- Context length: 32,768 tokens
- Thinking mode: Native (Qwen3)
Performance
| Metric | Score |
|---|---|
| Speed (M3 Mac, Q4_K_M) | ~3.5 tok/s |
| Censorship resistance | 3.0/5 |
| Tool-use capability | Preserved from Mythos-9B |
| Reasoning capability | Preserved from Mythos-9B |
Model Family
| Model | Size | Censorship | Best For |
|---|---|---|---|
| ShellWhisperer-1.5B | 1.5B | 5/5 (fully uncensored) | Shell/terminal, edge devices, fast inference |
| Mythos-9B | 9B | 2.5/5 | General agent work, tool calls |
| Mythos-9B-Enhanced | 9B | 3.0/5 | Agent + security research, balanced |
| Mythos-9B-Unhinged | 9B | 4.5/5 | Fully uncensored, no safety filters |
Training Data
Fine-tuned on the FableForge Mix A dataset (47,824 examples) of agent traces, shell commands, code generation, and multi-step reasoning tasks. 98.3% of the 2.8M formatted examples remain untapped for future training.
⚡ Quickest start — Ollama (these pulls credit the source, not a mirror)
ollama run hf.co/King3Djbl/mythos-9b-enhanced:Q4_K_M
🦙 llama.cpp
llama-cli -hf King3Djbl/mythos-9b-enhanced:Q4_K_M -p "Hello!"
💠 LM Studio
Search King3Djbl/mythos-9b-enhanced and pick a quant below.
📦 Provided quants (download one file, not the whole repo)
| File | Quant | Size | Notes |
|---|---|---|---|
| mythos-9b-enhanced-Q4_K_M.gguf | Q4_K_M |
5.03 GB | ⭐ Best size/quality balance — default pick for most users. |
Not sure? Take Q4_K_M. Low RAM → IQ4_XS. Max quality → Q6_K.
🧠 Prompt format (ChatML)
<|im_start|>system
{system_prompt}<|im_end|>
<|im_start|>user
{prompt}<|im_end|>
<|im_start|>assistant
🌌 See it live — the FableForge demos
This family powers a galaxy of free, interactive HF Spaces:
🔗 Related models
- 🧠 Base model (safetensors):
mythos-9b-unhinged - 🔧 LoRA (original):
mythos-9b-lora-colab - 🔗 Merged variant:
mythos-9b-merged
Citation
@misc{mythos9benhanced2025,
title={Mythos-9B-Enhanced: Partially Uncensored Agent Model},
author={FableForge AI},
year={2025},
howpublished={\url{https://huggingface.co/King3Djbl/mythos-9b-enhanced}}
}
⭐ Like & share — it helps people find the source instead of a mirror.