---
language: [en, zh, ja, ko, es, ru]
library_name: gguf
tags:
  - gguf
  - q4_k_m
  - quantized
  - qwen
  - qwopus
base_model: Jackrong/Qwopus3.6-27B-v2
license: apache-2.0
---

# Qwopus3.6-27B-v2 Q4_K_M GGUF

Q4_K_M quantization of [`Jackrong/Qwopus3.6-27B-v2`](https://huggingface.co/Jackrong/Qwopus3.6-27B-v2) via stock `llama.cpp` (ggml-org/llama.cpp), built locally with Metal support on Apple Silicon.

Preserves the MTP draft heads from the source — usable with speculative-decoding-enabled llama.cpp builds.

## Use

```bash
# llama.cpp / Ollama / LM Studio / koboldcpp — any stock-llama.cpp consumer
llama-cli -m Qwopus3.6-27B-v2-Q4_K_M.gguf -p "Hi"
```

Author's temperature recommendation: 0.75–1.0.

## Credits
- Source: [`Jackrong/Qwopus3.6-27B-v2`](https://huggingface.co/Jackrong/Qwopus3.6-27B-v2) by Jackrong
- GGUF conversion by [zaydiscold](https://huggingface.co/zaydiscold)

---

*Mix brought by the NOTORIOUS MLX.*