What are the hardware requirements for Marin 8B Instruct?

Marin 8B Instruct requires approximately 32.1 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 Marin 8B Instruct and how large is it?

Marin 8B Instruct is developed by marin-community, 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 Marin 8B Instruct?

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

Marin 8B Instruct by marin-community — VRAM 32.1GB, 4K context

Name: Marin 8B Instruct
Author: marin-community

Marin 8B Instruct is an open-source language model by marin-community. Features: 8b LLM, VRAM: 32.1GB, Context: 4K, License: apache-2.0, Instruction-Based, LLM Explorer Score: 0.2.

Arxiv:1909.09436 Arxiv:2109.01652 Arxiv:2110.14168 Arxiv:2302.13971 Arxiv:2309.12284 Arxiv:2310.03731 Arxiv:2410.05192 Arxiv:2412.02595 Base model:finetune:marin-comm... Base model:marin-community/mar... Conversational Dataset:bespokelabs/bespoke-st... Dataset:cognitivecomputations/... Dataset:facebook/natural reaso... Dataset:huggingfacetb/smoltalk Dataset:open-r1/openthoughts-1... Dataset:tiger-lab/acecode-87k Dataset:tuenguyen/dolphin r1 r... En Instruct Llama Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/marin-community/marin-8b-instruct

Marin 8B Instruct Parameters and Internals

LLM Name	Marin 8B Instruct
Repository 🤗	https://huggingface.co/marin-community/marin-8b-instruct
Base Model(s)	Marin 8B Base marin-community/marin-8b-base
Model Size	8b
Required VRAM	32.1 GB
Updated	2026-07-27
Maintainer	marin-community
Model Type	llama
Instruction-Based	Yes
Model Files	9.8 GB: 1-of-4 9.9 GB: 2-of-4 8.2 GB: 3-of-4 4.2 GB: 4-of-4
Supported Languages	en
Model Architecture	LlamaForCausalLM
License	apache-2.0
Context Length	4096
Model Max Length	4096
Transformers Version	4.51.3
Vocabulary Size	128256

Rank the Marin 8B Instruct 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 Marin 8B Instruct

Best Alternatives	Context / RAM	Downloads	Likes
...otron 8B UltraLong 4M Instruct	4192K / 32.1 GB	1135	125
UltraLong Thinking	4192K / 16.1 GB	2	3
...a 3.1 8B UltraLong 4M Instruct	4192K / 32.1 GB	176	24
...a 3.1 8B UltraLong 2M Instruct	2096K / 32.1 GB	875	9
...otron 8B UltraLong 2M Instruct	2096K / 32.1 GB	124	18
Cthulhu 8B V1.4	1048K / 16.1 GB	10	10
...raLong 1M Instruct Abliterated	1048K / 32.1 GB	4	9
...a 3.1 8B UltraLong 1M Instruct	1048K / 32.1 GB	1387	29
...otron 8B UltraLong 1M Instruct	1048K / 32.1 GB	702	59
Zero Llama 3.1 8B Beta6	1048K / 16.1 GB	7	1

Note: green Score (e.g. "73.2") means that the model is better than marin-community/marin-8b-instruct.