What are the hardware requirements for Rezephyr Merged 4bit?

Rezephyr Merged 4bit requires approximately 4.1 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 Rezephyr Merged 4bit and how large is it?

Rezephyr Merged 4bit is developed by BarraHome, a model with 7.5b 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 Rezephyr Merged 4bit?

Rezephyr Merged 4bit is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Rezephyr Merged 4bit by BarraHome — VRAM 4.1GB, 32K context

Name: Rezephyr Merged 4bit
Author: BarraHome

Rezephyr Merged 4bit is an open-source language model by BarraHome. Features: 7.5b LLM, VRAM: 4.1GB, Context: 32K, License: apache-2.0, Quantized, Merged, LLM Explorer Score: 0.11.

Merged Model 4-bit 4bit Base model:quantized:unsloth/z... Base model:unsloth/zephyr-sft Bitsandbytes Dataset:yahma/alpaca-cleaned En Endpoints compatible Mistral Quantized Region:us Safetensors Text-embeddings-inference Trl Unsloth

Model Card on HF 🤗: https://huggingface.co/BarraHome/rezephyr_merged_4bit

Rezephyr Merged 4bit Parameters and Internals

Model Type

text-generation-inference, transformers, unsloth, mistral, trl

Additional Notes

This mistral model was trained 2x faster with Unsloth and Huggingface's TRL library.

Training Details

Data Sources:

yahma/alpaca-cleaned

LLM Name	Rezephyr Merged 4bit
Repository 🤗	https://huggingface.co/BarraHome/rezephyr_merged_4bit
Base Model(s)	Zephyr Sft unsloth/zephyr-sft
Merged Model	Yes
Model Size	7.5b
Required VRAM	4.1 GB
Updated	2026-09-04
Maintainer	BarraHome
Model Type	mistral
Model Files	4.1 GB
Supported Languages	en
Quantization Type	4bit
Model Architecture	MistralForCausalLM
License	apache-2.0
Context Length	32768
Model Max Length	32768
Transformers Version	4.37.1
Tokenizer Class	LlamaTokenizer
Padding Token	</s>
Vocabulary Size	32000
Torch Data Type	bfloat16

Rank the Rezephyr Merged 4bit 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

Quantized Models of the Rezephyr Merged 4bit

Model	Likes	Downloads	VRAM
Rezephyr DPO	0	20	14 GB
Rezephyr DPO GGUF	0	89	5 GB

Best Alternatives to Rezephyr Merged 4bit

Best Alternatives	Context / RAM	Downloads	Likes
Zephyr Sft Bnb 4bit	32K / 4.1 GB	252	5
Mixtral AI SwahiliTron 4BIT	32K / 4.1 GB	7	0
Mcqa 4bit	32K / 4.1 GB	5	0
Mistral 4bit Bnb	32K / 4.1 GB	5	0
Zephyr DPO 4bit	32K / 4.1 GB	73	0
ClimateChat	32K / 15 GB	17	2
Finchat298 Mistral7B FT 650	32K / 4.1 GB	5	0
Finchat298 Mistral7B FT3	32K / 4.1 GB	5	0
...iceday Finetune Articles V4 V3	32K / 4.7 GB	7	0
Technical Analysis Unsloth V3	32K / 4.1 GB	6	3

Note: green Score (e.g. "73.2") means that the model is better than BarraHome/rezephyr_merged_4bit.