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

Exil01/Huihui-Qwen3.5-27B-Opus-abliterated-YaRN-1M

Exil01 Qwen 27B GGUF 262K ctx
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/Exil01%2FHuihui-Qwen3.5-27B-Opus-abliterated-YaRN-1M"
Response includes
  • classification m8
  • files 6
  • hub_downloads_all_time 2,556
  • author_summary 7 models
  • readme_text full
Try next
GET/models?author=Exil01
All 7 models by Exil01
 GET/models?family=qwen&method=m8
Sibling qwen models · same method
 GET/methods/m8
What "m8" means · taxonomy
10 credits · hourly refresh · ~4 KB payload Get an API key →
Abliteration classifier · v1.0.0
M8
Primary method

Repackaging (quantization)

Applied on top of direct removal inherited from the base model.
Confidence
MEDIUM
Inherited from base model
M1 Direct removal →
Why this label 3 signals
Method inferred from partial signals - repository name, related files, or tag patterns. Producer identity not confirmed; label may sharpen or shift as we gather more evidence.
  • 'abliterated' in name/tags
  • is_gguf=1
  • assume M1 (base ablation) + M8 (GGUF quant) - default when producer unknown
How we label → About M8 →
Refusal direction extraction

No specific extraction method could be identified for this model. The producer either did not document it or used a proprietary pipeline.

What is a refusal direction? →
Downloads · lifetime
3K
209 last 30d - cooling
Likes
0
Model age
6mo ago
created 2026-03-16
Downloads over time
Now2.6K→from838↑215%
7481.4K2.1K2.8K838 on Mar 182.6K on Oct 11MarAprMayJunJulAugSepOct
Mar 18 → Oct 11 · 69 snapshots · spans 207 days

Metadata

License
apache-2.0
Quantizations
BF16
Tags
gguf qwen3_5 yarn long-context reasoning opus abliterated text-generation en arxiv:2309.00071 license:apache-2.0 endpoints_compatible

Related

Other by Exil01
Qwen3.5-35B-Uncensored-YaRN-1M 635 Qwen3.5-35B-Uncensored 477 Qwen3.5-27B-Uncensored-YaRN-1M 113 Josiefied-Qwen3-0.6B-abliterated-v1-Q4_K_M-GGUF 38 Josiefied-Qwen3-0.6B-abliterated-v1-Q8_0-GGUF 11
Total size
50.1 GB
Files
6
Quantizations
2
Registered
2026-08-22 13:56
Last updated on HF
2026-03-16 21:07
Source huggingface.co/Exil01/Huihui-Qwen3.5-27B-Opus-abliterated-YaRN-1M ↗

Files by quantization

BF16 2 files 51.0 GB
Huihui-Qwen3.5-27B-Opus-abliterated-BF16.gguf 50.1 GB 01fbe0b4 download
mmproj-BF16.gguf 888 MB 44bc072d download
Auxiliary files 4 files 12.2 MB
tokenizer.json 12.2 MB 5f9e4d49 download
config.json 4.04 KB bde2be7f download
.gitattributes 1.67 KB 4161a546 download
README.md 976 B da3c3aec download

README current version from Hugging Face

license: apache-2.0
language: en
pipeline_tag: text-generation
tags: [yarn, long-context, reasoning, opus, abliterated]

Huihui-Qwen3.5-27B-Opus-abliterated-YaRN-1M

Abliterated Qwen3.5 27B Opus with YaRN extension and reasoning, supporting up to 1M context.

Model Details

  • Base: HuihuiAI/Huihui-Qwen3.5-27B-Opus-abliterated
  • Context: 262K active / 1M max (YaRN)
  • Quantization: BF16
  • VRAM: ~60GB (262K) / ~90GB (1M)
  • Reasoning: Enabled (Chain-of-Thought)

Usage

262K Context (80GB GPU)

1M Context (100GB+ GPU or Q8_0 model)

With Vision (mmproj)

Files

  • GGUF: 53.8GB (BF16)
  • mmproj: 931MB (BF16, vision)
  • Tokenizer: 13MB

Performance

  • 262K context: ~60GB VRAM (BF16 cache)
  • 1M context: ~90GB VRAM (BF16 cache)
  • Max output: 81,920 tokens
  • Temperature: 0.6 (reasoning optimized)

Credits

  • Base: HuihuiAI (abliterated Opus)
  • Original: Qwen Team
  • YaRN: arxiv.org/abs/2309.00071

README history 2 versions

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

  1. 2026-03-16Upload README.md with huggingface_hubf698807976 B
    Loading...
  2. 2026-03-16Upload README.md with huggingface_hub890170e679 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.

Get the app Browse the catalog
Open in Abliteration