QWEN-3.8 27B Abliterated GGUF PRIVATE METHOD DONE IN 30 MIN

:broken_chain:DONE WITH PRIVATE METHOD! ABLATION DONE IN 30 MINUTES!

:wrapped_gift: A smart, private AI that runs on your own computer — no login, no monthly fee, works offline, and it never says “I can’t help with that.”
It’s the normal Qwen with the refusals stripped out, so it stays just as smart — passed 14 of 15 skill tests. Free to keep (Apache-2.0).

:hammer_and_wrench: Get the free app to run it → LM Studio (easiest — Ollama or Jan work too). Install it, load the file, start chatting.

Which size? Bigger = smarter, needs a stronger PC (decent GPU or lots of RAM):

  • :green_circle: 10.7 GB — lightest, modest PC
  • :blue_circle: 16.5 GB — the sweet spot, grab this
  • :purple_circle: 22.1 GB — near-perfect, strong PC

:backhand_index_pointing_right: Download it here

10 Likes

I don’t think this is MoE (Mixture of Experts) model.

is the 10gb enough for coding?

Lower quants didnt do as good on mmlu benchmarks id use bigger quants

i am looking for qwen uncensored model could you please help me with that

This is uncencored