What are the hardware requirements for Marco Mini Base?

Marco Mini Base requires approximately 34.8 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 Marco Mini Base and how large is it?

Marco Mini Base is developed by ATH-MaaS, a model with 100b 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 Marco Mini Base?

Marco Mini Base is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Marco Mini Base by ATH-MaaS — VRAM 34.8GB, 32K context

Name: Marco Mini Base
Author: ATH-MaaS

Marco Mini Base is an open-source language model by ATH-MaaS. Features: 100b LLM, VRAM: 34.8GB, Context: 32K, License: apache-2.0, LLM Explorer Score: 0.24.

Ar Az Bn Conversational Cs Dataset:allenai/dolma3 dolmino... Dataset:allenai/dolmino-mix-11... Dataset:huggingfacefw/fineweb-... Dataset:huggingfacetb/finemath Dataset:llm360/megamath Dataset:nvidia/nemotron-cc-mat... Dataset:nvidia/nemotron-cc-v2 Dataset:nvidia/nemotron-cc-v2.... Dataset:nvidia/nemotron-pretra... Dataset:nvidia/nemotron-pretra... Dataset:nvidia/openmathinstruc... Dataset:open-thoughts/openthou... Dataset:opencsg/fineweb-edu-ch... De El En Endpoints compatible Es Fr He Hu Id It Ja Kk Ko Mixture-of-experts Moe Ms Multilingual Ne Nl Pl Pt Qwen3 moe Region:us Ro Ru Safetensors Sharded Tensorflow Th Tr Uk Upcycling Ur Vi Zh

Model Card on HF 🤗: https://huggingface.co/ATH-MaaS/Marco-Mini-Base

Marco Mini Base Parameters and Internals

LLM Name	Marco Mini Base
Repository 🤗	https://huggingface.co/ATH-MaaS/Marco-Mini-Base
Model Size	100b
Required VRAM	34.8 GB
Updated	2026-08-27
Maintainer	ATH-MaaS
Model Type	qwen3_moe
Model Files	2.0 GB: 1-of-18 2.0 GB: 2-of-18 2.0 GB: 3-of-18 2.0 GB: 4-of-18 2.0 GB: 5-of-18 2.0 GB: 6-of-18 2.0 GB: 7-of-18 2.0 GB: 8-of-18 2.0 GB: 9-of-18 2.0 GB: 10-of-18 2.0 GB: 11-of-18 2.0 GB: 12-of-18 2.0 GB: 13-of-18 2.0 GB: 14-of-18 2.0 GB: 15-of-18 2.0 GB: 16-of-18 2.0 GB: 17-of-18 0.8 GB: 18-of-18
Supported Languages	en zh ar de es fr ko ja pt tr id it nl pl ru vi th he uk ms bn cs ur kk el ro hu ne az
Model Architecture	Qwen3MoeForCausalLM
License	apache-2.0
Context Length	32768
Model Max Length	32768
Transformers Version	4.57.1
Tokenizer Class	Qwen2Tokenizer
Padding Token	<\|endoftext\|>
Vocabulary Size	151936
Errors	replace

Rank the Marco Mini Base 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 Marco Mini Base

Best Alternatives	Context / RAM	Downloads	Likes
Marco Mini Base	32K / 34.8 GB	21	8