← back to catalog · registered 2026-11-15 13:03

brainworkup/Qwen3.8-27B-Ultimate-Heretic-Uncensored-oQ6e-mtp

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/brainworkup%2FQwen3.8-27B-Ultimate-Heretic-Uncensored-oQ6e-mtp"
Response includes
  • classification m3
  • files 16
  • author_summary 10 models
  • readme_text full
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
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)
Refusal direction extracted via
Extraction technique

Difference-of-means

Confidence
HIGH
Why we say so
name contains 'heretic'; Heretic default extraction is difference-of-means (Arditi 2024)
Downloads · 30-day
119
Likes
0
Model age
6d ago
created 2026-11-15
Downloads over time
Now 119 → from 0 ↑ 0%
04487131 0 on Nov 15 119 on Nov 22 Nov
Nov 15 → Nov 22 · 8 snapshots · spans 7 days

Metadata

Tags
mlx safetensors qwen3_5 oq quantized mtp omlx oQ6e base_model:DavidAU/Qwen3.8-27B-Ultimate-Heretic-Uncensored base_model:quantized:DavidAU/Qwen3.8-27B-Ultimate-Heretic-Uncensored 6-bit region:us

Related

Total size
22.2 GB
Files
16
Quantizations
1
Registered
2026-11-15 13:03
Last updated on HF
2026-11-22 05:54

Files by quantization

Auxiliary files 16 files 22.3 GB
model-00004-of-00005.safetensors 4.69 GB 51a1c84b download
model-00001-of-00005.safetensors 4.68 GB e264c30e download
model-00002-of-00005.safetensors 4.68 GB 23be4f60 download
model-00003-of-00005.safetensors 4.68 GB d035f2b5 download
model-00005-of-00005.safetensors 3.50 GB 5ff54d85 download
tokenizer.json 19.1 MB 6f32ce20 download
vocab.json 6.41 MB 0aa0ce06 download
chat_template.jinja 246 KB 485b7fc6 download
model.safetensors.index.json 207 KB 7bdb4a1f download
oq_imatrix_report.json 30.5 KB 64993fa8 download
config.json 15.7 KB 20758609 download
tokenizer_config.json 8.48 KB 1a7e62d0 download
.gitattributes 1.53 KB 52373fe2 download
README.md 982 B 70ae1138 download
preprocessor_config.json 390 B 2ea84a43 download
generation_config.json 214 B 0bc3addd download

README current version from Hugging Face


library_name: mlx
tags:

  • mlx
  • oq
  • quantized
  • mtp

Qwen3.8-27B-Ultimate-Heretic-Uncensored-oQ6e-mtp

This model was quantized using oQ (oMLX v0.8.0rc1) mixed-precision quantization.

The effective average is 6.872 bits/weight for the complete checkpoint.

Quantization and Bit Distribution

Storage format Logical weights Share of total Tensor storage Effective bits/weight
Affine 6-bit, group 64 24.543B 88.345% 18.572 GiB 6.50
Affine 8-bit, group 64 2.722B 9.798% 2.694 GiB 8.50
BF16 0.516B 1.857% 0.961 GiB 16.00
Total 27.781B 100% 22.226 GiB 6.872

Storage Breakdown

Component Logical weights Storage
Language model 26.896B 22.526 GB / 20.979 GiB
Vision 0.461B 0.921 GB / 0.858 GiB
MTP head 0.425B 0.418 GB / 0.389 GiB
Total 27.781B 23.865 GB / 22.226 GiB

README history 4 versions

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

  1. 2026-11-22 updated title ced9f77 6.2 KB
    Loading...
  2. 2026-11-15 formatted d08ce12 6.2 KB
    Loading...
  3. 2026-11-15 added details of model 58abadc 6.9 KB
    Loading...
  4. 2026-11-15 Upload Qwen3.8-27B-Ultimate-Heretic-Uncensored-oQ6e-mtp via oMLX 4c7cc93 982 B
    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.

Open in Abliteration