What are the hardware requirements for StagAI CyberSecurity 26B A4B V2 Q4 MLX?

StagAI CyberSecurity 26B A4B V2 Q4 MLX requires approximately 35.4 GB of VRAM. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed StagAI CyberSecurity 26B A4B V2 Q4 MLX and how large is it?

StagAI CyberSecurity 26B A4B V2 Q4 MLX is developed by deer-sec, a model with 26b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

How does StagAI CyberSecurity 26B A4B V2 Q4 MLX perform on standard benchmarks?

StagAI CyberSecurity 26B A4B V2 Q4 MLX has the following published scores: LMSYS ELO 1438. Compare against reference models on this page or on the LLM Explorer leaderboards.

StagAI CyberSecurity 26B A4B V2 Q4 MLX by deer-sec — VRAM 35.4GB

Name: StagAI CyberSecurity 26B A4B V2 Q4 MLX
Author: deer-sec

StagAI CyberSecurity 26B A4B V2 Q4 MLX is an open-source language model by deer-sec. Features: 26b LLM, VRAM: 35.4GB, License: apache-2.0, Quantized, ELO: 1438.

4-bit 4bit Conversational Cyber Cybersecurity Dataset:alicankiraz0/cybersecu... Dataset:deer-sec/deer sec-japa... En Finetuned Gemma Gemma4 Ja Mlx Multilingual Q4 Quantized Reasoning Region:us Safetensors Security Sharded Tensorflow Zh

Model Card on HF 🤗: https://huggingface.co/deer-sec/StagAI-CyberSecurity-26B-A4B-v2-Q4-MLX

StagAI CyberSecurity 26B A4B V2 Q4 MLX Benchmarks

LMSys ELO 1438 vs 1272 (so35) 13.1%

^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)

StagAI CyberSecurity 26B A4B V2 Q4 MLX Parameters and Internals

LLM Name	StagAI CyberSecurity 26B A4B V2 Q4 MLX
Repository 🤗	https://huggingface.co/deer-sec/StagAI-CyberSecurity-26B-A4B-v2-Q4-MLX
Base Model(s)	google/gemma-4-26b-a4b google/gemma-4-26b-a4b
Model Size	26b
Required VRAM	35.4 GB
Updated	2026-10-17
Maintainer	deer-sec
Model Type	gemma4
Model Files	5.3 GB: 1-of-8 5.1 GB: 2-of-8 4.9 GB: 3-of-8 4.9 GB: 4-of-8 4.9 GB: 5-of-8 4.9 GB: 6-of-8 4.9 GB: 7-of-8 0.5 GB: 8-of-8
Supported Languages	en ja zh
Quantization Type	q4
Model Architecture	Gemma4ForConditionalGeneration
License	apache-2.0
Transformers Version	5.5.0.dev0
Tokenizer Class	GemmaTokenizer
Padding Token	<pad>

Rank the StagAI CyberSecurity 26B A4B V2 Q4 MLX 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 StagAI CyberSecurity 26B A4B V2 Q4 MLX

Best Alternatives	Context / RAM	Downloads	Likes
...6B A4B It Qat Q4 0 Unquantized	0K / 51.6 GB	33696	50
Gemma 4 26B A4b It 4bit	0K / 15.4 GB	27543	79
...6B A4B It Qat Q4 0 Unquantized	0K / 51.6 GB	3487	3
OTel 2.0 Decision 26B A4B 4bit	0K / 17.3 GB	63	1
Gemma 4 26B A4b It 8bit	0K / 27.9 GB	2917	14
Gemma 4 26B A4b It UD MLX 4bit	0K / 16.2 GB	1527	34
Gemma 4 26B A4b It MLX 8bit	0K / 27.9 GB	936	16
...at Q4 0 Unquantized Heretic Ja	0K / 51.6 GB	34	1
Gemma 4 26B A4b 4bit	0K / 15.6 GB	488	19
Gemma 4 26B A4b It 6bit	0K / 21.7 GB	611	7

Note: green Score (e.g. "73.2") means that the model is better than deer-sec/StagAI-CyberSecurity-26B-A4B-v2-Q4-MLX.