LLM EXPLORER 58,873 MODELS INDEXED

Qwen3.8 27B ABLITERATED EXL3 2.75bpw by WatchDG

By WatchDG · 39 downloads
HF Repo ↗ Copy URL

Qwen3.8 27B ABLITERATED EXL3 2.75bpw is an open-source language model by WatchDG. Features: 27b LLM, VRAM: 13GB, License: apache-2.0, LLM Explorer Score: 0.29.

  2.75bpw   27b   Abliterated Base model:blackfrost-ai/qwen3... Base model:quantized:blackfros...   Conversational   En   Exl3   Image-text-to-text   Quantized   Qwen3   Qwen3.8   Qwen3 5   Region:us   Safetensors   Sharded   Tensorflow   Vision   Vision-language
Model Card on HF 🤗: https://huggingface.co/WatchDG/Qwen3.8-27B-ABLITERATED-EXL3-2.75bpw 

Qwen3.8 27B ABLITERATED EXL3 2.75bpw Parameters and Internals

LLM NameQwen3.8 27B ABLITERATED EXL3 2.75bpw
Repository 🤗https://huggingface.co/WatchDG/Qwen3.8-27B-ABLITERATED-EXL3-2.75bpw 
Base Model(s)  Blackfrost-AI/Qwen3.8-27B-ABLITERATED-BF16   Blackfrost-AI/Qwen3.8-27B-ABLITERATED-BF16
Model Size27b
Required VRAM13 GB
Updated2026-08-18
MaintainerWatchDG
Model Typeqwen3_5
Model Files  8.5 GB: 1-of-2   4.5 GB: 2-of-2
Supported Languagesen
Model ArchitectureQwen3_5ForConditionalGeneration
Licenseapache-2.0
Model Max Length262144
Transformers Version5.8.0.dev0
Tokenizer ClassQwen2Tokenizer
Padding Token<|endoftext|>
Errorsreplace

Best Alternatives to Qwen3.8 27B ABLITERATED EXL3 2.75bpw

Best Alternatives
Context / RAM
Downloads
Likes
Qwen3.8 27B FP80K / 0.4 GB123157416
Qwen3.6 27B0K / 55.4 GB67281382197
Qwen3.8 27B0K / 55.6 GB28753
Qwen3.6 27B NVFP40K / 22 GB1887525419
Huihui Qwen3.8 27B Abliterated0K / 55.6 GB148576
ThinkingCap Qwen3.6 27B0K / 54.7 GB35947614
Qwen3.6 27B FP80K / 0.4 GB7740472342
...n3.8 27B MTPLX Optimized Speed0K / 19.6 GB698950
Qwen3.5 27B0K / 55.2 GB25362941028
Qwen3.5 27B0K / 55.2 GB811516
Note: green Score (e.g. "73.2") means that the model is better than WatchDG/Qwen3.8-27B-ABLITERATED-EXL3-2.75bpw.