LLM EXPLORER 66,978 MODELS INDEXED

Qwen3.8 4B Distill ARA Abliterated by Austriani

By Austriani · 33 downloads
HF Repo ↗ Copy URL

Qwen3.8 4B Distill ARA Abliterated is an open-source language model by Austriani. Features: 4b LLM, VRAM: 9.1GB, License: apache-2.0, LLM Explorer Score: 0.3.

  Abliterated   Ara   Arbitrary-rank-ablation   Austriani Base model:empero-ai/qwen3.8-4... Base model:finetune:empero-ai/...   Chat   Conversational   Creative-writing   Decensored   Distill   Empero-ai   En   Endpoints compatible   Heretic   Image-text-to-text   Qwen3.8   Qwen3.8-4b   Qwen3 5   Refusal-removed   Region:us   Roleplay   Safetensors   Sharded   Tensorflow   Uncensored
Model Card on HF 🤗: https://huggingface.co/Austriani/Qwen3.8-4B-Distill-ARA-Abliterated 

Qwen3.8 4B Distill ARA Abliterated Parameters and Internals

LLM NameQwen3.8 4B Distill ARA Abliterated
Repository 🤗https://huggingface.co/Austriani/Qwen3.8-4B-Distill-ARA-Abliterated 
Base Model(s)  empero-ai/Qwen3.8-4B-Distill   empero-ai/Qwen3.8-4B-Distill
Model Size4b
Required VRAM9.1 GB
Updated2026-11-09
MaintainerAustriani
Model Typeqwen3_5
Model Files  5.0 GB: 1-of-2   4.1 GB: 2-of-2
Supported Languagesen
Model ArchitectureQwen3_5ForConditionalGeneration
Licenseapache-2.0
Model Max Length262144
Transformers Version5.22.1
Tokenizer ClassQwen3_5Tokenizer
Padding Token<|endoftext|>
Errorsreplace

Best Alternatives to Qwen3.8 4B Distill ARA Abliterated

Best Alternatives
Context / RAM
Downloads
Likes
AgentHorse 4B256K / 8.4 GB9737
Qwen3.5 4B0K / 9.3 GB6575502801
NuExtract30K / 0.2 GB441570324
LightOnOCR 3 4B0K / 9.1 GB239571
Agents A1 4B0K / 9.1 GB17757872
...en3.5 4B 3D Coder MTP Repacked0K / 9.3 GB21328
...5 4B GLB 3D Coder MTP Repacked0K / 17.7 GB25571
FrogNano 4B 26090K / 9.3 GB733113
RuBio 4B0K / 9.3 GB38810
Qwen3.5 4B Base0K / 9.3 GB22781479
Note: green Score (e.g. "73.2") means that the model is better than Austriani/Qwen3.8-4B-Distill-ARA-Abliterated.