What are the hardware requirements for Nethena 20B Glued GPTQ?

Nethena 20B Glued GPTQ requires approximately 10.5 GB of VRAM and supports a context window of 4K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Nethena 20B Glued GPTQ and how large is it?

Nethena 20B Glued GPTQ is developed by TheBloke, a model with 20b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

Where can I download or evaluate Nethena 20B Glued GPTQ?

Nethena 20B Glued GPTQ is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Nethena 20B Glued GPTQ by TheBloke — VRAM 10.5GB, 4K context

Name: Nethena 20B Glued GPTQ
Author: TheBloke

Nethena 20B Glued GPTQ is an open-source language model by TheBloke. Features: 20b LLM, VRAM: 10.5GB, Context: 4K, License: cc-by-nc-4.0, Quantized, LLM Explorer Score: 0.1.

4-bit Base model:athirdpath/nethena-... Base model:quantized:athirdpat... Gptq Llama Quantized Region:us Safetensors

Model Card on HF 🤗: https://huggingface.co/TheBloke/Nethena-20B-Glued-GPTQ

Nethena 20B Glued GPTQ Parameters and Internals

Model Type

llama

Additional Notes

This model is a test, exploring the effects of 'gluing' the components of the 20b model together to reduce iconic word replacement errors, increase lucidity, and improve recall. Trained on NSFW-oriented content, it passes subjective testing with flying colors.

Training Details

Data Sources:

Medical texts, Excerpts from short stories and novellas, PIPPA, GPT-4 generated data

Data Volume:

~500k tokens

Methodology:

Alpaca formatted with LORA application

Context Length:

2048

Model Architecture:

LORA-augmented, NSFW-oriented, aligned to reduce iconic word replacement errors

Input Output

Input Format:

Alpaca-style prompts

LLM Name	Nethena 20B Glued GPTQ
Repository 🤗	https://huggingface.co/TheBloke/Nethena-20B-Glued-GPTQ
Model Name	Nethena 20B Glued
Model Creator	Raven
Base Model(s)	Nethena 20B Glued athirdpath/Nethena-20b-Glued
Model Size	20b
Required VRAM	10.5 GB
Updated	2026-08-02
Maintainer	TheBloke
Model Type	llama
Model Files	10.5 GB
GPTQ Quantization	Yes
Quantization Type	gptq
Model Architecture	LlamaForCausalLM
License	cc-by-nc-4.0
Context Length	4096
Model Max Length	4096
Transformers Version	4.35.2
Tokenizer Class	LlamaTokenizer
Padding Token	[PAD]
Vocabulary Size	32001
Torch Data Type	float16

Rank the Nethena 20B Glued GPTQ capabilities

Have you tried this model? Rate its performance — this feedback helps the ML community find the right model for their needs.

Instruction Following and Task Automation
Factuality and Completeness of Knowledge
Censorship and Alignment
Data Analysis and Insight Generation
Text Generation
Text Summarization and Feature Extraction
Code Generation
Multi-Language Support and Translation

Best Alternatives to Nethena 20B Glued GPTQ

Best Alternatives	Context / RAM	Downloads	Likes
Norocetacean 20B 10K GPTQ	10K / 10.5 GB	9	2
Orca2myth7.2 GPTQ	4K / 10.9 GB	7	4
DaringMaid 20B GPTQ	4K / 10.5 GB	5	5
Rose Kimiko 20B GPTQ	4K / 10.5 GB	0	1
Iambe Storyteller 20B GPTQ	4K / 10.5 GB	4	2
Iambe 20B DARE GPTQ	4K / 10.5 GB	4	1
Noromaid 20B V0.1.1 GPTQ	4K / 10.5 GB	11	8
Rose 20B GPTQ	4K / 10.5 GB	4	3
Nethena 20B GPTQ	4K / 10.5 GB	12	7
MXLewd L2 20B GPTQ	4K / 10.5 GB	8	9

Note: green Score (e.g. "73.2") means that the model is better than TheBloke/Nethena-20B-Glued-GPTQ.