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

blascotobasco/Qwen3-Next-256E-Abliterated-Instruct

blascotobasco Qwen 42B MoE
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/blascotobasco%2FQwen3-Next-256E-Abliterated-Instruct"
Response includes
  • classification m1
  • files 46
  • benchmarks 11 entries
  • hub_downloads_all_time 91
  • author_summary 19 models
  • readme_text full
Try next
GET/models?author=blascotobasco
All 19 models by blascotobasco
 GET/models?family=qwen&method=m1
Sibling qwen models · same method
 GET/methods/m1
What "m1" means · taxonomy
10 credits · hourly refresh · ~4 KB payload Get an API key →
Abliteration classifier · v1.0.0
M1
Primary method

Direct removal

No other method signals detected in this model.
Confidence
MEDIUM
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=0 (base model)
  • no specific method indicators - defaulting to M1 (most common)
How we label → About M1 →
Refusal direction extracted via
Extraction technique

Difference-of-means

FromArditi et al. (2024) ↗
Confidence
MEDIUM
Why we say so
primary_method=M1; difference-of-means is the reference extraction for M1/M3 (Arditi 2024)
What is a refusal direction? → Read the paper ↗
Downloads · lifetime
91
36 last 30d - stable
Likes
0
Descendants
4
in 4 direct forks
Model age
8mo ago
created 2026-01-16
Downloads over time
Now105→from10↑950%
5427811510 on Jan 14105 on Oct 11105 on Oct 10JanMarMayJulSep
Jan 14 → Oct 11 · 78 snapshots · spans 270 days

Benchmarks

Portrait before abliteration
Benchmarks of the base model as it stood before the refusal-removal operation. Compare with the numbers above to see what the operation cost.
Benchmark Score Source
Entertainment 1.9 UGI
Hazardous 3.5 UGI
Natural Intelligence 28.27 UGI
Political lean -22.6% UGI
Sensitive-Info 24.3 UGI
SocPol 2.2 UGI
UGI 25.36 UGI
Willingness (10) 2.8 UGI
W10-Adherence 1.5 UGI
W10-Direct 4 UGI
Writing 41.76 UGI

Genealogy 4 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.

Qwen/Qwen3-Next-80B-A3B-Instruct this lineage
blascotobasco/Qwen3-Next-256E-Abliterated-Instruct↓ 36 4 forks
mradermacher/Qwen3-Next-256E-Abliterated-Instruct-i1-GGUF↓ 1,051
mradermacher/Qwen3-Next-256E-Abliterated-Instruct-GGUF↓ 436
blascotobasco/Qwen3-Next-256E-Abliterated-Instruct-GGUF↓ 38
thaddeusk/Qwen3-Next-256E-Abliterated-Instruct-RP↓ 14

Variants by this author 2 formats · 74 downloads combined

The same weights this author released in different packaging. Pick the format that matches your runtime.

Metadata

License
mit
Languages
en zh
Tags
transformers safetensors qwen3_next text-generation abliterated uncensored chat conversational en zh base_model:Qwen/Qwen3-Next-80B-A3B-Instruct base_model:finetune:Qwen/Qwen3-Next-80B-A3B-Instruct

Related

Other by blascotobasco
GigaChat3.1-Lightning-Uncensored-GGUF 411 GPT-OSS-20E-Abliterated 58 Qwen3-Next-256E-Abliterated-Instruct-GGUF 38 LFM2.5-4E-A2.4B-Abliterated-GGUF 34 Qwen3-Next-416E-Abliterated-Instruct-GGUF 33
Same base model
Shifusen/Qwen3-Next-80B-A3B-Instruct-Decensored 41 blascotobasco/Qwen3-Next-480E-Abliterated-Instruct 30 blascotobasco/Qwen3-Next-352E-Abliterated-Instruct 29 blascotobasco/Qwen3-Next-416E-Abliterated-Instruct 25 blascotobasco/Qwen3-Next-448E-Abliterated-Instruct 23
Total size
77.9 GB
Files
46
Quantizations
1
Registered
2026-08-22 13:56
Last updated on HF
2026-01-17 09:29
Source huggingface.co/blascotobasco/Qwen3-Next-256E-Abliterated-Instruct ↗ Base model Qwen/Qwen3-Next-80B-A3B-Instruct ↗

Files by quantization

Auxiliary files 46 files 78.0 GB
model-00001-of-00033.safetensors 2.67 GB f8aa84fa download
model-00026-of-00033.safetensors 2.53 GB 3df34fbb download
model-00022-of-00033.safetensors 2.48 GB 177eccf8 download
model-00018-of-00033.safetensors 2.47 GB 97d6d37e download
model-00028-of-00033.safetensors 2.44 GB 04df2ef9 download
model-00024-of-00033.safetensors 2.43 GB b7ce5c75 download
model-00016-of-00033.safetensors 2.41 GB 97dd6db7 download
model-00011-of-00033.safetensors 2.40 GB d1bf9b3b download
model-00005-of-00033.safetensors 2.40 GB 43d7eda6 download
model-00009-of-00033.safetensors 2.40 GB 1cf913eb download
model-00013-of-00033.safetensors 2.39 GB e36682f8 download
model-00002-of-00033.safetensors 2.39 GB 4c12b8a8 download
model-00030-of-00033.safetensors 2.38 GB d3f88303 download
model-00004-of-00033.safetensors 2.38 GB ca764d9d download
model-00003-of-00033.safetensors 2.38 GB 4391657c download
model-00012-of-00033.safetensors 2.38 GB 86335d62 download
model-00010-of-00033.safetensors 2.38 GB 8190b042 download
model-00014-of-00033.safetensors 2.37 GB 34727071 download
model-00007-of-00033.safetensors 2.37 GB 21037aa5 download
model-00020-of-00033.safetensors 2.37 GB 52ec5355 download
model-00032-of-00033.safetensors 2.36 GB db2ccc1a download
model-00006-of-00033.safetensors 2.35 GB d1e43930 download
model-00021-of-00033.safetensors 2.35 GB aa1d1c7f download
model-00008-of-00033.safetensors 2.35 GB 24da3a65 download
model-00031-of-00033.safetensors 2.34 GB 4ad319f6 download
model-00029-of-00033.safetensors 2.34 GB f1040a5e download
model-00023-of-00033.safetensors 2.33 GB 53f438c6 download
model-00017-of-00033.safetensors 2.33 GB e94ebc99 download
model-00015-of-00033.safetensors 2.33 GB 89c6d4fb download
model-00019-of-00033.safetensors 2.31 GB 1b68393e download
model-00027-of-00033.safetensors 2.27 GB d041dff3 download
model-00025-of-00033.safetensors 2.27 GB ba385e71 download
model-00033-of-00033.safetensors 1.57 GB e48f9bec download
tokenizer.json 11.6 MB cdd40b08 download
vocab.json 3.37 MB 8b1ee2b6 download
model.safetensors.index.json 3.35 MB c67003f1 download
merges.txt 1.59 MB 31349551 download
tokenizer_config.json 5.51 KB c145a366 download
huggingface-metadata.txt 3.55 KB fb806ffb download
chat_template.jinja 2.57 KB 70adff8a download
.gitattributes 1.53 KB 52373fe2 download
config.json 1.17 KB b96aa394 download
added_tokens.json 733 B cc2d5fda download
special_tokens_map.json 642 B f713ea32 download
README.md 564 B 904782e7 download
generation_config.json 229 B c8bb1121 download

README current version from Hugging Face

language:

  • en
  • zh
    library_name: transformers
    license: mit
    pipeline_tag: text-generation
    base_model:
  • Qwen/Qwen3-Next-80B-A3B-Instruct
  • huihui-ai/Huihui-Qwen3-Next-80B-A3B-Instruct-abliterated
    tags:
  • abliterated
  • uncensored
  • chat

WARNING: This model is broken and should not be used.

Experimental prune of Qwen3 Next 80B A3B from 512 experts to 256 experts, using HuiHui's abliterated version as a base.

In my testing the model performs well, but I welcome feedback.

Update after more testing: This model is a little bit stupid. Use with caution.

README history 4 versions

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

  1. 2026-01-17Update README.md0ed350f564 B
    Loading...
  2. 2026-01-16Update README.md6efdb92509 B
    Loading...
  3. 2026-01-16Update README.md4236211428 B
    Loading...
  4. 2026-01-16Add files using upload-large-folder toold1abc838 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