Jimster480/Deck-Opus-NEO-CODE-HERE-2T-OT-Q6_K-MTP-GGUF overview
Based on: https://huggingface.co/DavidAU/Qwen3.6 40B Claude 4.6 Opus Deckard Heretic Uncensored Thinking NEO CODE Di IMatrix MAX GGUF Found this model to be be…
Runs locally from ~884.6 MB disk (4 GB VRAM class GPUs with llama.cpp / guIDE).
Repository Files & Downloads
Model Details
| Model ID | Jimster480/Deck-Opus-NEO-CODE-HERE-2T-OT-Q6_K-MTP-GGUF |
|---|---|
| Author | Jimster480 |
| Pipeline | — |
| License | apache-2.0 |
| Base model | DavidAU/Qwen3.6-40B-Claude-4.6-Opus-Deckard-Heretic-Uncensored-Thinking-NEO-CODE-Di-IMatrix-MAX-GGUF |
| Last modified | 2026-07-28T06:50:26.000Z |
Model README
---
license: apache-2.0
base_model:
- >-
DavidAU/Qwen3.6-40B-Claude-4.6-Opus-Deckard-Heretic-Uncensored-Thinking-NEO-CODE-Di-IMatrix-MAX-GGUF
tags:
- unsloth
- fine tune
- heretic
- uncensored
- abliterated
- multi-stage tuned.
- all use cases
- coder
- creative
- creative writing
- fiction writing
- MTP
---
Based on:
https://huggingface.co/DavidAU/Qwen3.6-40B-Claude-4.6-Opus-Deckard-Heretic-Uncensored-Thinking-NEO-CODE-Di-IMatrix-MAX-GGUF
Found this model to be better than Qwen3.6-27B Unsloth UD-Q8_K_XL so I worked to add MTP to it.
MTP Grafted from Unsloth MTP header
- 50-100% Speed Increase in Decode vs Original
The version with -PT in it is using a Post Trained MTP head created by another community member from DavidAU's model community page.
It sometimes offers additional performance over the standard MTP head.
Run Jimster480/Deck-Opus-NEO-CODE-HERE-2T-OT-Q6_K-MTP-GGUF with guIDE
Download guIDE — the AI-native code editor with local LLM inference and 69 built-in tools.
Source: Hugging Face · Compare models