What are the hardware requirements for AegisX1 8B A1B?

AegisX1 8B A1B requires approximately 0.7 GB of VRAM. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed AegisX1 8B A1B and how large is it?

AegisX1 8B A1B is developed by Lobus22, a model with 8b 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 AegisX1 8B A1B?

AegisX1 8B A1B is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

AegisX1 8B A1B by Lobus22 — VRAM 0.7GB

Name: AegisX1 8B A1B
Author: Lobus22

AegisX1 8B A1B is an open-source language model by Lobus22. Features: 8b LLM, VRAM: 0.7GB, Quantized, LLM Explorer Score: 0.29.

Arxiv:1910.09700 4bit Adapter Base model:adapter:unsloth/met... Base model:unsloth/meta-llama-... Finetuned Lora Peft Quantized Region:us Safetensors Sft Trl Unsloth

Model Card on HF 🤗: https://huggingface.co/Lobus22/AegisX1-8B-A1B

AegisX1 8B A1B Parameters and Internals

LLM Name	AegisX1 8B A1B
Repository 🤗	https://huggingface.co/Lobus22/AegisX1-8B-A1B
Base Model(s)	unsloth/Meta-Llama-3.1-8B-bnb-4bit unsloth/Meta-Llama-3.1-8B-bnb-4bit
Model Size	8b
Required VRAM	0.7 GB
Updated	2026-09-16
Maintainer	Lobus22
Model Files	0.7 GB
Quantization Type	4bit
Model Architecture	Adapter
Model Max Length	131072
Is Biased	none
Tokenizer Class	TokenizersBackend
Padding Token	<\|finetune_right_pad_id\|>
PEFT Type	LORA
LoRA Model	Yes
PEFT Target Modules	v_proj\|gate_proj\|o_proj\|k_proj\|up_proj\|q_proj\|down_proj
LoRA Alpha	64
LoRA Dropout	0
R Param	64

Rank the AegisX1 8B A1B 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 AegisX1 8B A1B

Best Alternatives	Context / RAM	Downloads	Likes
Soro	0K / 5.8 GB	1107	0
Llama 3.1 8B PubMedQA QLoRA	0K / 0.2 GB	6	2
Arabic DeepSeek R1 Distill 8B	0K / 0 GB	9	4
Llama 3 8B Muxue 1.0 Lora	0K / 0.2 GB	5	0
...ma3 Instruct Chat FCAIBylaw V1	0K / 0.2 GB	5	0
BlueMoon Llama3	0K / 0.7 GB	7	0
Theory Of Mind Llama3	0K / 1.3 GB	7	0
RP Format QuoteAsterisk Llama3	0K / 1.3 GB	6	0
Llama3 RP ORPO LoRA	0K / 0.7 GB	5	0
NoWarning Llama3	0K / 1.3 GB	5	0