What are the hardware requirements for MAIR?

MAIR requires approximately 15.2 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 MAIR and how large is it?

MAIR is developed by Daemontatox, 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 MAIR?

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

MAIR by Daemontatox — VRAM 15.2GB, 32K context

Name: MAIR
Author: Daemontatox

MAIR is an open-source language model by Daemontatox. Features: 7b LLM, VRAM: 15.2GB, Context: 32K, License: apache-2.0, Quantized, Instruction-Based, LLM Explorer Score: 0.16.

4bit Autotrain compatible Base model:finetune:qwen/qwen2... Base model:qwen/qwen2.5-7b-ins... Chain-of-thought Conversational Dataset:daemontatox/alpaca rea... Dataset:daemontatox/longcot-re... En Endpoints compatible Instruct Long-cot-reasoning Mamba2 Pytorch Quantized Qwen2 Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/critical-hf/MAIR

MAIR Parameters and Internals

LLM Name	MAIR
Repository 🤗	https://huggingface.co/critical-hf/MAIR
Base Model(s)	Qwen/Qwen2.5-7B-Instruct Qwen/Qwen2.5-7B-Instruct
Model Size	7b
Required VRAM	15.2 GB
Updated	2025-04-11
Maintainer	Daemontatox
Model Type	qwen2
Instruction-Based	Yes
Model Files	4.9 GB: 1-of-4 4.9 GB: 2-of-4 4.3 GB: 3-of-4 1.1 GB: 4-of-4 4.9 GB: 1-of-4 4.9 GB: 2-of-4 4.3 GB: 3-of-4 1.1 GB: 4-of-4
Supported Languages	en
Quantization Type	4bit
Model Architecture	Qwen2ForCausalLM
License	apache-2.0
Context Length	32768
Model Max Length	32768
Transformers Version	4.46.3
Tokenizer Class	Qwen2Tokenizer
Padding Token	<\|PAD_TOKEN\|>
Vocabulary Size	152064
Torch Data Type	float16
Errors	replace

Rank the MAIR 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 MAIR

Best Alternatives	Context / RAM	Downloads	Likes
Qwen2.5 7B Instruct 1M 4bit	986K / 4.3 GB	444	11
...wen2.5 7B Instruct 1M Bnb 4bit	986K / 5.5 GB	267	1
Qwen2.5 7B Instruct 1M 8bit	986K / 8.1 GB	63	4
Qwen2.5 7B Instruct 1M 6bit	986K / 6.2 GB	52	2
...B Instruct 1M Unsloth Bnb 4bit	986K / 7.5 GB	37	3
Legal Qwen2.5 7B Sft	128K / 15.2 GB	155	0
Alisia 7B Instruct V1	128K / 15.4 GB	114	2
...5 7B Instruct SFT Meme LoRA V7	128K / 15.2 GB	6	0
Gte Qwen2 7B Instruct 4bit DWQ	128K / 4.3 GB	88	3
Qwen2.5 7B Legal Chatbot Exp1	32K / 15.2 GB	290	0

Note: green Score (e.g. "73.2") means that the model is better than critical-hf/MAIR.