This model is for research and educational purposes only. It is trained on a dataset that includes profanity, slurs, and edgy internet culture. It is not intended to be used as a production system, and its only current value is entertainment.

However, care has been taken to retain the model's original capabilities for long-horizon or complex requests.

This model will likely underperform when benchmarked against the base model.

Available Quantizations

Format	Use Case
`Q8_0`	Near-lossless, good balance of quality and size
`Q6_K`	High quality with moderate compression
`Q5_K_M`	Good quality with stronger compression
`Q4_K_M`	Maximum compression, acceptable quality loss

Usage

Compatible with llama.cpp and any GGUF-compatible inference engine (e.g., LM Studio, Ollama, text-generation-webui).

bash

# Example with llama.cpp
./llama-cli -m Qwopus-3.5-27B-V3-Stabbed-Q4_K_M.gguf -p "Your prompt here"

Stabhappy/Qwopus-3.5-27B-V3-Stabbed-GGUF

Qwopus 3.5 27B V3 "Stabbed" — GGUF

Disclaimer

Available Quantizations

Usage

Stabhappy/
Qwopus-3.5-27B-V3-Stabbed-GGUF