---
license: other
language:
- en
tags:
- gguf
- llama.cpp
- llama-3.3
- 70b
- 128k
- roleplay
- uncensored
model_type: llama
base_model: ReadyArt/L3.3-The-Omega-Directive-70B-Unslop-v2.0
---

# Loki-Omega-70B-GGUF (Q6_K)

This repo hosts **GGUF** split weights for running *ReadyArt/L3.3-The-Omega-Directive-70B-Unslop-v2.0* with **llama.cpp** (Q6_K quant).
Files:
- `L3.3-The-Omega-Directive-70B-Unslop-v2.0.Q6_K.gguf.part1of2`
- `L3.3-The-Omega-Directive-70B-Unslop-v2.0.Q6_K.gguf.part2of2`

> Note: HF serves split files as a single artifact to supported clients (e.g., `huggingface_hub`, `llama-cpp-python`). Keep both parts in the repo.

## Quick use (llama.cpp server, OpenAI-compatible)
```bash
pip install "llama-cpp-python[server]"
python -m llama_cpp.server \
  --model /path/to/L3.3-The-Omega-Directive-70B-Unslop-v2.0.Q6_K.gguf \
  --host 0.0.0.0 --port 8000 \
  --n_ctx 32000

Quant: Q6_K (keeps nuance well).

Intended for RP/system-prompt use; bring your own safety/persona rules.

## Warning
You can use it if you want, but he vomits on everything.  
Loki loves being told: **“NO LOKI! NO!”**