What are the hardware requirements for Notus 7B V1.8.0bpw H8 EXL2?

Notus 7B V1.8.0bpw H8 EXL2 requires approximately 7.4 GB of VRAM and supports a context window of 32K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Notus 7B V1.8.0bpw H8 EXL2 and how large is it?

Notus 7B V1.8.0bpw H8 EXL2 is developed by LoneStriker, a model with 7b 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 Notus 7B V1.8.0bpw H8 EXL2?

Notus 7B V1.8.0bpw H8 EXL2 is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Notus 7B V1.8.0bpw H8 EXL2 by LoneStriker — VRAM 7.4GB, 32K context

Name: Notus 7B V1.8.0bpw H8 EXL2
Author: LoneStriker

Notus 7B V1.8.0bpw H8 EXL2 is an open-source language model by LoneStriker. Features: 7b LLM, VRAM: 7.4GB, Context: 32K, License: mit, Quantized, LLM Explorer Score: 0.1.

Base model:alignment-handbook/... Base model:finetune:alignment-... Conversational Dataset:argilla/ultrafeedback-... Dpo En Endpoints compatible Exl2 Mistral Model-index Preference Quantized Region:us Rlaif Safetensors Ultrafeedback

Model Card on HF 🤗: https://huggingface.co/LoneStriker/notus-7b-v1-8.0bpw-h8-exl2

Notus 7B V1.8.0bpw H8 EXL2 Parameters and Internals

Model Type

GPT-like 7B model DPO fine-tuned

Additional Notes

Notus is fine-tuned using Direct Preference Optimization (DPO) over the base model `zephyr-7b-sft-full`. It uses a curarily fixed version of the UltraFeedback dataset to train the model.

Supported Languages

en (Mainly English)

Training Details

Data Sources:

openbmb/UltraFeedback

Hardware Used:

8 x A100 40GB

Model Architecture:

GPT-like

LLM Name	Notus 7B V1.8.0bpw H8 EXL2
Repository 🤗	https://huggingface.co/LoneStriker/notus-7b-v1-8.0bpw-h8-exl2
Base Model(s)	Zephyr 7B Sft Full alignment-handbook/zephyr-7b-sft-full
Model Size	7b
Required VRAM	7.4 GB
Updated	2026-07-30
Maintainer	LoneStriker
Model Type	mistral
Model Files	7.4 GB 0.0 GB
Supported Languages	en
Quantization Type	exl2
Model Architecture	MistralForCausalLM
License	mit
Context Length	32768
Model Max Length	32768
Transformers Version	4.35.0
Tokenizer Class	LlamaTokenizer
Padding Token	</s>
Vocabulary Size	32000
Torch Data Type	bfloat16

Rank the Notus 7B V1.8.0bpw H8 EXL2 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 Notus 7B V1.8.0bpw H8 EXL2

Best Alternatives	Context / RAM	Downloads	Likes
...t 3.5 0106 128K 4.0bpw H6 EXL2	128K / 3.9 GB	4	1
...t 3.5 0106 128K 8.0bpw H8 EXL2	128K / 7.4 GB	1	1
...aude Sonnet Mistral 7B It V0.3	32K / 14.5 GB	7	2
Mistral 7B Instruct V0.2 Fp16	32K / 14.4 GB	26	0
Mistral 7B V0.3 Bnb 4bit	32K / 4.1 GB	381323	22
Mistral 7B Instruct V0.2 4bit	32K / 4.3 GB	79	1
...tral 7B Instruct V0.3 Bnb 4bit	32K / 4.1 GB	34633	42
Mistral 7B Instruct V0.2 8bit	32K / 7.6 GB	5	1
...tral 7B Instruct V0.2 Bnb 4bit	32K / 4.1 GB	7160	35
Mixtral V0.3 Full 16bit	32K / 14.5 GB	12	0

Note: green Score (e.g. "73.2") means that the model is better than LoneStriker/notus-7b-v1-8.0bpw-h8-exl2.