LLM EXPLORER 62,076 MODELS INDEXED

MK4 Ginko V2 by MK4-Research

By MK4-Research · 409 downloads
HF Repo ↗ Copy URL

MK4 Ginko V2 is an open-source language model by MK4-Research. Features: 27b LLM, VRAM: 21.8GB, License: apache-2.0, Quantized, LLM Explorer Score: 0.32.

  4bit   6-bit Base model:mlx-community/qwen3... Base model:quantized:mlx-commu...   Conversational   Cybersecurity   En   Guard-analysis   Mlx   Quantized   Qwen3.8   Qwen3 5   Region:us   Safetensors   Security-code-review   Sharded   Tensorflow
Model Card on HF 🤗: https://huggingface.co/MK4-Research/MK4-Ginko-v2 

MK4 Ginko V2 Parameters and Internals

LLM NameMK4 Ginko V2
Repository 🤗https://huggingface.co/MK4-Research/MK4-Ginko-v2 
Base Model(s)  mlx-community/Qwen3.8-27B-4bit   mlx-community/Qwen3.8-27B-4bit
Model Size27b
Required VRAM21.8 GB
Updated2026-09-24
MaintainerMK4-Research
Model Typeqwen3_5
Model Files  5.4 GB: 1-of-5   5.3 GB: 2-of-5   5.3 GB: 3-of-5   4.8 GB: 4-of-5   1.0 GB: 5-of-5
Supported Languagesen
Quantization Type4bit
Model ArchitectureQwen3_5ForConditionalGeneration
Licenseapache-2.0
Model Max Length262144
Transformers Version5.8.0.dev0
Tokenizer ClassQwen2Tokenizer
Padding Token<|endoftext|>
Errorsreplace

Best Alternatives to MK4 Ginko V2

Best Alternatives
Context / RAM
Downloads
Likes
Qwen3.6 27B OptiQ 4bit256K / 18.7 GB1306559
Bonsai 27B Mlx 1bit0K / 5.1 GB809309205
Ternary Bonsai 27B Mlx 2bit0K / 8.5 GB809024168
Qwen3.8 27B MLX 5bit0K / 19.5 GB46750442
...onsai 2 27B CRACK Ternary JANG0K / 8.1 GB268913
Qwen3.6 27B MLX 8bit0K / 29.4 GB87914514
Qwen3.6 27B MLX 6bit0K / 22.7 GB8129772
Qwen3.8 27B 8bit0K / 29.4 GB1389233
Qwen3.6 27B MLX 4bit0K / 16 GB84686116
Ternary Bonsai 2 27B Vllm0K / 10.3 GB20484
Note: green Score (e.g. "73.2") means that the model is better than MK4-Research/MK4-Ginko-v2.