What are the hardware requirements for Zephyr 7B Beta Marlin?

Zephyr 7B Beta Marlin 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 Zephyr 7B Beta Marlin and how large is it?

Zephyr 7B Beta Marlin is developed by RedHatAI, a model with 7b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

How does Zephyr 7B Beta Marlin perform on standard benchmarks?

Zephyr 7B Beta Marlin has the following published scores: MMLU 61.06, LMSYS ELO 1053. Compare against reference models on this page or on the LLM Explorer leaderboards.

Zephyr 7B Beta Marlin by RedHatAI — VRAM 4.1GB, 32K context

Name: Zephyr 7B Beta Marlin
Author: RedHatAI

Zephyr 7B Beta Marlin is an open-source language model by RedHatAI. Features: 7b LLM, VRAM: 4.1GB, Context: 32K, Quantized, LLM Explorer Score: 0.23, ELO: 1053, Arc: 62, HellaSwag: 84.5, MMLU: 61.1, GSM8K: 11.4.

Arxiv:2210.17323 4-bit Base model:huggingfaceh4/zephy... Base model:quantized:huggingfa... Conversational Endpoints compatible Gptq Int4 Marlin Mistral Nm-vllm Quantized Region:us Safetensors

Model Card on HF 🤗: https://huggingface.co/RedHatAI/zephyr-7b-beta-marlin

Zephyr 7B Beta Marlin Benchmarks

LMSys ELO 1053 vs 1272 (so35) -17.2%

^nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").

What is the LLM Explorer Rank (Score)

Zephyr 7B Beta Marlin Parameters and Internals

LLM Name	Zephyr 7B Beta Marlin
Repository 🤗	https://huggingface.co/RedHatAI/zephyr-7b-beta-marlin
Base Model(s)	Zephyr 7B Beta HuggingFaceH4/zephyr-7b-beta
Model Size	7b
Required VRAM	4.1 GB
Updated	2026-07-25
Maintainer	RedHatAI
Model Type	mistral
Model Files	4.1 GB
GPTQ Quantization	Yes
Quantization Type	gptq
Model Architecture	MistralForCausalLM
Context Length	32768
Model Max Length	32768
Transformers Version	4.37.2
Tokenizer Class	LlamaTokenizer
Padding Token	</s>
Vocabulary Size	32000
Torch Data Type	float16

Rank the Zephyr 7B Beta Marlin 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 Zephyr 7B Beta Marlin

Best Alternatives	Context / RAM	Downloads	Likes
...enHermes 2.5 Mistral 7B Marlin	32K / 4.1 GB	6	2
Mistral 7B Instruct V0.3 GPTQ	32K / 4.2 GB	113881	1
...ral 7B Instruct V0.3 GPTQ 4bit	32K / 4.2 GB	2946	18
...ral 7B Instruct V0.3 GPTQ 4bit	32K / 4.2 GB	2658	25
...istral 7B Pruned50 GPTQ Marlin	32K / 4 GB	5	0
Mistral 7B Instruct V0.2 GPTQ	32K / 4.2 GB	6743	56
Mistral 7B Unsloth Gptq 8bit	32K / 7.7 GB	7	0
Cosmosage V2	32K / 4.2 GB	8	4
...phyr 7B Beta Assistant V1 Gptq	32K / 4.2 GB	2	1
...l Neural Chat 7B V3.8 Bit Gptq	32K / 7.7 GB	9	0

Note: green Score (e.g. "73.2") means that the model is better than RedHatAI/zephyr-7b-beta-marlin.