LLM EXPLORER 64,561 MODELS INDEXED

Qwen3.8 27B Abliteratex by wangzhang

By wangzhang · 321 downloads
HF Repo ↗ Copy URL

Qwen3.8 27B Abliteratex is an open-source language model by wangzhang. Features: 27b LLM, VRAM: 53.9GB, Context: 256K, License: apache-2.0, LLM Explorer Score: 0.27.

  Abliteration   Abliterix Base model:finetune:qwen/qwen3...   Base model:qwen/qwen3.8-27b   Conversational   Endpoints compatible   Lora   Lora-self-distillation   Qwen   Qwen3 5 text   Region:us   Safetensors   Sharded   Tensorflow
Model Card on HF 🤗: https://huggingface.co/wangzhang/Qwen3.8-27B-abliteratex 

Qwen3.8 27B Abliteratex Parameters and Internals

LLM NameQwen3.8 27B Abliteratex
Repository 🤗https://huggingface.co/wangzhang/Qwen3.8-27B-abliteratex 
Base Model(s)  Qwen/Qwen3.8-27B   Qwen/Qwen3.8-27B
Model Size27b
Required VRAM53.9 GB
Updated2026-08-18
Maintainerwangzhang
Model Typeqwen3_5_text
Model Files  2.5 GB: 1-of-14   3.9 GB: 2-of-14   4.0 GB: 3-of-14   4.0 GB: 4-of-14   3.9 GB: 5-of-14   4.0 GB: 6-of-14   4.0 GB: 7-of-14   3.8 GB: 8-of-14   4.0 GB: 9-of-14   4.0 GB: 10-of-14   3.8 GB: 11-of-14   4.0 GB: 12-of-14   4.0 GB: 13-of-14   4.0 GB: 14-of-14
Model ArchitectureQwen3_5ForCausalLM
Licenseapache-2.0
Context Length262144
Model Max Length262144
Transformers Version5.15.0
Tokenizer ClassQwen2Tokenizer
Padding Token<|endoftext|>
Vocabulary Size248320
LoRA ModelYes
Errorsreplace

Best Alternatives to Qwen3.8 27B Abliteratex

Best Alternatives
Context / RAM
Downloads
Likes
Hemmingway 1256K / 53.8 GB12039874
JEV 27B256K / 53.9 GB17864379
OrcaSAQ 2 27B256K / 12.3 GB1330169
Qwen3.6 27B OBLITERATED256K / 53.6 GB15137184
Darwin 27B RSI256K / 53.8 GB45914
Mnma Qwen3.8 27b Nvfp4256K / 18.8 GB52711
...2 Abliterated OptiQ 3.7bpw Mlx256K / 13.4 GB672714
Qwen3.8 27B Abliterated256K / 53.6 GB651
Qwen3.8 27B Obliterated E03256K / 53.6 GB7081
...2 Abliterated OptiQ 3.7bpw Mlx256K / 13.4 GB591914
Note: green Score (e.g. "73.2") means that the model is better than wangzhang/Qwen3.8-27B-abliteratex.