---

license: apache-2.0

base_model:

• markmonger/Qwen3.6-40B-Claude-4.6-Opus-Deckard-Heretic-Uncensored-Thinking-iMatrix-MTP-GGUF

tags:

• gguf
• qwen3.6
• qwen
• mtp
• speculative-decoding
• llama.cpp
• q4_k_m
• quantized

pipeline_tag: text-generation

---

Qwen3.6 40B Heretic MTP GGUF Q4_K_M

This repository provides a Q4_K_M GGUF quantization derived from:

markmonger/Qwen3.6-40B-Claude-4.6-Opus-Deckard-Heretic-Uncensored-Thinking-iMatrix-MTP-GGUF

The source repository provides the MTP GGUF model as q8_0-v1.gguf and is licensed under Apache-2.0.

Files

| File | Quantization | SHA256 |

| --- | --- | --- |

| q4_k_m-v1.gguf | Q4_K_M | c9791ad1fda0c2cb0290c268c035d888121e3d5e84a9e6f0a6b34618cc081d90 |

Usage

With llama.cpp:

llama-cli -hf m233zhan/Qwen3.6-40B-Heretic-MTP-GGUF-Q4_K_M:q4_k_m-v1.gguf

Or start an OpenAI-compatible local server:

llama-server -hf m233zhan/Qwen3.6-40B-Heretic-MTP-GGUF-Q4_K_M:q4_k_m-v1.gguf

Provenance

Model lineage shown by the upstream repository:

• Base model: Qwen/Qwen3.6-27B
• Finetuned model: DavidAU/Qwen3.6-40B-Claude-4.6-Opus-Deckard-Heretic-Uncensored-Thinking
• MTP GGUF source: markmonger/Qwen3.6-40B-Claude-4.6-Opus-Deckard-Heretic-Uncensored-Thinking-iMatrix-MTP-GGUF

This repository only adds the Q4_K_M GGUF quantization.