LLM EXPLORER 59,420 MODELS INDEXED

Qwen3.8 27B Huihui Abliterated OQ4e MTP MLX by KostkaIT

By KostkaIT · 447 downloads
HF Repo ↗ Copy URL

Qwen3.8 27B Huihui Abliterated OQ4e MTP MLX is an open-source language model by KostkaIT. Features: 27b LLM, VRAM: 16.9GB, License: apache-2.0, LLM Explorer Score: 0.32.

  4-bit   Abliterated   Apple-silicon Base model:huihui-ai/huihui-qw... Base model:quantized:huihui-ai...   Conversational   Image-text-to-text   Imatrix   Mlx   Mtp   Native-mtp   Omlx   Oq4e   Quantized   Qwen   Qwen3   Qwen3.8   Qwen3 5   Region:us   Safetensors   Sharded   Tensorflow   Vision
Model Card on HF 🤗: https://huggingface.co/KostkaIT/Qwen3.8-27B-Huihui-Abliterated-oQ4e-MTP-MLX 

Qwen3.8 27B Huihui Abliterated OQ4e MTP MLX Parameters and Internals

LLM NameQwen3.8 27B Huihui Abliterated OQ4e MTP MLX
Repository 🤗https://huggingface.co/KostkaIT/Qwen3.8-27B-Huihui-Abliterated-oQ4e-MTP-MLX 
Base Model(s)  huihui-ai/Huihui-Qwen3.8-27B-abliterated   huihui-ai/Huihui-Qwen3.8-27B-abliterated
Model Size27b
Required VRAM16.9 GB
Updated2026-08-24
MaintainerKostkaIT
Model Typeqwen3_5
Model Files  5.0 GB: 1-of-4   5.0 GB: 2-of-4   5.0 GB: 3-of-4   1.9 GB: 4-of-4
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 Huihui Abliterated OQ4e MTP MLX

Best Alternatives
Context / RAM
Downloads
Likes
Qwen3.8 27B MLX0K / 16 GB141118
Qwen3.8 27B FP80K / 0.4 GB123157416
Qwen3.8 27B0K / 55.6 GB272721
Qwen3.6 27B0K / 55.4 GB67281382197
Qwen3.8 27B0K / 55.6 GB28753
Qwen3.6 27B NVFP40K / 22 GB1887525419
ThinkingCap Qwen3.6 27B0K / 54.7 GB35947614
Qwen3.6 27B FP80K / 0.4 GB7740472342
Qwen3.5 27B0K / 55.2 GB25362941028
Qwen3.8 27B Escha W20K / 10.1 GB56197
Note: green Score (e.g. "73.2") means that the model is better than KostkaIT/Qwen3.8-27B-Huihui-Abliterated-oQ4e-MTP-MLX.