Get the app
← back to catalog · registered 2026-08-22 13:56

mlx-community/Qwen3.5-40B-Claude-4.6-Opus-Deckard-Heretic-Uncensored-Thinking-4.5bit-msq

mlx-community Qwen 5.4B multimodal second-order
Your rig guess connected
? Why do I need an app?
Reading your rig…

This is a rough estimate. Install the free app - we'll show exact numbers.

Reading real hardware from your app right now. Numbers below are exact.

Below is the per-quantization compatibility for this model. 

curl -H "Authorization: Bearer $ABL_KEY" \
     "https://abliteration.org/api/v1/models/mlx-community%2FQwen3.5-40B-Claude-4.6-Opus-Deckard-Heretic-Uncensored-Thinking-4.5bit-msq"
Response includes
  • classification m3
  • files 19
  • hub_downloads_all_time 352
  • author_summary 207 models
  • readme_text full
Try next
GET/models?author=mlx-community
All 207 models by mlx-community
 GET/models?family=qwen&method=m3
Sibling qwen models · same method
 GET/methods/m3
What "m3" means · taxonomy
10 credits · hourly refresh · ~4 KB payload Get an API key →
Abliteration classifier · v1.0.0
M3
Primary method

Layer-wise ablation

Applied on top of direct removal inherited from the base model.
Confidence
HIGH
Inherited from base model
M1 Direct removal →
Why this label 2 signals
Producer identity confirmed by naming conventions, tags or the model card. This label is very unlikely to change.
  • 'heretic' in model name (Heretic-produced)
  • Heretic uses layer-wise optimization (M3) with underlying direction removal (M1)
How we label → About M3 →
Refusal direction extracted via
Extraction technique

Difference-of-means

FromArditi et al. (2024) ↗
Confidence
HIGH
Why we say so
name contains 'heretic'; Heretic default extraction is difference-of-means (Arditi 2024)
What is a refusal direction? → Read the paper ↗
Downloads · lifetime
352
62 last 30d - stable
Likes
0
Model age
5mo ago
created 2026-04-18

Training datasets

Corpora the author lists in the model card. Datasets tracked in our /datasets catalog carry a category badge linking to the workflow stage. Others open on Hugging Face.

TeichAI/claude-4.5-opus-high-reasoning-250x DavidAU/PkDick-Deckard-5-Datasets
Downloads over time
Now379→from47↑706%
3015828541247 on Apr 22379 on Oct 11AprMayJunJulAugSepOct
Apr 22 → Oct 11 · 64 snapshots · spans 172 days

Genealogy 0 direct forks

Full fork graph →

This model's place in the market. Above: what it was derived from. Below: the tree of everything derived from it.

DavidAU/Qwen3.5-27B-Deckard-PKD-Heretic-Uncensored-Thinking↓ 87 this lineage
mlx-community/Qwen3.5-40B-Claude-4.6-Opus-Deckard-Heretic-Uncensored-Thinking-4.5bit-msq↓ 62

Metadata

License
apache-2.0
Languages
en zh
Tags
transformers safetensors qwen3_5 image-text-to-text unsloth fine tune heretic uncensored abliterated multi-stage tuned. all use cases coder

Related

Other by mlx-community
Qwen3.8-27B-Uncensored-OptiQ-4bit 16K Qwen3.6-40B-Claude-4.6-Opus-Deckard-Heretic-Uncensored-Thinking-8bit 7K Qwen3.8-Flash-Next-Uncensored-oQ5e-mtp 5K Josiefied-Qwen3-4B-abliterated-v1-4bit 4K Huihui-Qwen3.6-35B-A3B-Claude-4.7-Opus-abliterated-mlx-8bit 4K
Same base model
mradermacher/Qwen3.5-27B-Deckard-PKD-Heretic-Uncensored-Thinking-i1-GGUF 2K mradermacher/Qwen3.5-27B-Deckard-PKD-Heretic-Uncensored-Thinking-GGUF 585 DavidAU/Qwen3.5-40B-Claude-4.6-Opus-Deckard-Heretic-Uncensored-Thinking 270 DavidAU/Qwen3.5-21B-Claude-4.6-Opus-Deckard-Heretic-Uncensored-Thinking 159 Bercraft/Qwen3.5-27B-Deckard-PKD-Heretic-Uncensored-Thinking-Q8_0-GGUF 149
Total size
22.7 GB
Files
19
Quantizations
1
Registered
2026-08-22 13:56
Last updated on HF
2026-04-18 18:48
Source huggingface.co/mlx-community/Qwen3.5-40B-Claude-4.6-Opus-Deckard-Heretic-Uncensored-Thinking-4.5bit-msq ↗ Base model DavidAU/Qwen3.5-27B-Deckard-PKD-Heretic-Uncensored-Thinking →

Files by quantization

Auxiliary files 19 files 22.8 GB
model-00002-of-00005.safetensors 5.00 GB 5d4c4a28 download
model-00004-of-00005.safetensors 4.99 GB 551dd56c download
model-00001-of-00005.safetensors 4.98 GB c3b3bb67 download
model-00003-of-00005.safetensors 4.97 GB 7ccc275c download
model-00005-of-00005.safetensors 2.79 GB 96b4aa35 download
tokenizer.json 19.1 MB 87a7830d download
vocab.json 6.41 MB 0aa0ce06 download
model.safetensors.index.json 320 KB ff9f30a3 download
config.json 80.6 KB d197423e download
chat_template.jinja 7.78 KB 5a663a52 download
.DS_Store 6.00 KB 5008ddfc download
.gitattributes 1.53 KB 52373fe2 download
README.md 1.36 KB d596208e download
processor_config.json 1.27 KB 8af8110f download
tokenizer_config.json 1.11 KB a068e246 download
quant_recipe.json 485 B 3a86b541 download
preprocessor_config.json 390 B 2ea84a43 download
video_preprocessor_config.json 385 B 3ba673a5 download
.gitignore 18.0 B 62687331 download

README current version from Hugging Face

language:

  • en
  • zh
    license: apache-2.0
    tags:
  • unsloth
  • fine tune
  • heretic
  • uncensored
  • abliterated
  • multi-stage tuned.
  • all use cases
  • coder
  • creative
  • creative writing
  • fiction writing
  • plot generation
  • sub-plot generation
  • fiction writing
  • story generation
  • scene continue
  • storytelling
  • fiction story
  • science fiction
  • romance
  • all genres
  • story
  • writing
  • vivid prosing
  • vivid writing
  • fiction
  • roleplaying
  • bfloat16
  • all use cases
    datasets:
  • TeichAI/claude-4.5-opus-high-reasoning-250x
  • DavidAU/PkDick-Deckard-5-Datasets
    base_model:
  • DavidAU/Qwen3.5-27B-Deckard-PKD-Heretic-Uncensored-Thinking
    pipeline_tag: image-text-to-text
    library_name: transformers

Qwen3.5-40B-Claude-4.6-Opus-Deckard-Heretic-Uncensored-Thinking — MLX 4.5 BPW

Mixed-precision MLX quantization of DavidAU/Qwen3.5-40B-Claude-4.6-Opus-Deckard-Heretic-Uncensored-Thinking, quantized with MLX Smart Quantize (MSQ) — my own sensitivity-based mixed-precision quantization method for Apple Silicon. It measures per-layer NMSE and assigns optimal bit widths automatically, combining architecture knowledge with measured data.

Details

  • Type: Vision (VLM)
  • Average: 4.5 bits per weight
  • Method: MLX Smart Quantize (MSQ)
  • AWQ scaling: applied to 120 groups

README history 1 version

The author's README evolved over time. Click a version to see its content at that point.

  1. 2026-04-18Add files using upload-large-folder tool407595e1.4 KB
    Loading...
Catalog is the map. Apps are the tools.

Run models on your own machine, not in the cloud.

Every model page has an "Open in Abliteration" button that hands the model directly to the first-party desktop client, at the quantization your rig can actually run. No API keys, no subscription, no prompt leakage.

Get the app Browse the catalog
Open in Abliteration