LLM EXPLORER 65,418 MODELS INDEXED

HuginnV5.5 12.6B 8.0bpw H8 EXL2 by LoneStriker

By LoneStriker · 3 downloads
HF Repo ↗ Copy URL

HuginnV5.5 12.6B 8.0bpw H8 EXL2 is an open-source language model by LoneStriker. Features: 12.6b LLM, VRAM: 13.1GB, Context: 32K, License: cc-by-4.0, Quantized, LLM Explorer Score: 0.1.

  Conversational   Endpoints compatible   Exl2   Mistral   Quantized   Region:us   Safetensors   Sharded   Tensorflow
Model Card on HF 🤗: https://huggingface.co/LoneStriker/HuginnV5.5-12.6B-8.0bpw-h8-exl2 

HuginnV5.5 12.6B 8.0bpw H8 EXL2 Parameters and Internals

Model Type 
text generation, conversational AI
Use Cases 
Areas:
conversational tasks, creative tasks
Applications:
story writing, roleplaying, helping writers with creative tasks
Primary Use Cases:
coming up with creative ideas
Additional Notes 
Huginn is an experimental frankenmerge model using multiple 7B models through the Dare-ties method.
Input Output 
Input Format:
alpaca for instructions, chatML format for conversation
LLM NameHuginnV5.5 12.6B 8.0bpw H8 EXL2
Repository 🤗https://huggingface.co/LoneStriker/HuginnV5.5-12.6B-8.0bpw-h8-exl2 
Model Size12.6b
Required VRAM13.1 GB
Updated2026-08-01
MaintainerLoneStriker
Model Typemistral
Model Files  8.6 GB: 1-of-2   4.5 GB: 2-of-2
Quantization Typeexl2
Model ArchitectureMistralForCausalLM
Licensecc-by-4.0
Context Length32768
Model Max Length32768
Transformers Version4.36.2
Tokenizer ClassLlamaTokenizer
Vocabulary Size32000
Torch Data Typefloat16

Best Alternatives to HuginnV5.5 12.6B 8.0bpw H8 EXL2

Best Alternatives
Context / RAM
Downloads
Likes
...al Nemo Instruct 2407 Bnb 4bit128K / 8.3 GB727731
...istral Nemo Base 2407 Bnb 4bit128K / 8.3 GB189815
HuginnV5.5 12.6B32K / 25.8 GB00
Note: green Score (e.g. "73.2") means that the model is better than LoneStriker/HuginnV5.5-12.6B-8.0bpw-h8-exl2.