What are the hardware requirements for Mistral Small 3.2 24B Instruct 2506 ARA Heretic?

Mistral Small 3.2 24B Instruct 2506 ARA Heretic requires approximately 48.2 GB of VRAM. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Mistral Small 3.2 24B Instruct 2506 ARA Heretic and how large is it?

Mistral Small 3.2 24B Instruct 2506 ARA Heretic is developed by richardyoung, a model with 24b 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 Mistral Small 3.2 24B Instruct 2506 ARA Heretic?

Mistral Small 3.2 24B Instruct 2506 ARA Heretic is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Mistral Small 3.2 24B Instruct 2506 ARA Heretic by richardyoung — VRAM 48.2GB

Name: Mistral Small 3.2 24B Instruct 2506 ARA Heretic
Author: richardyoung

Mistral Small 3.2 24B Instruct 2506 ARA Heretic is an open-source language model by richardyoung. Features: 24b LLM, VRAM: 48.2GB, License: apache-2.0, Instruction-Based, LLM Explorer Score: 0.3.

Abliterated Ar Base model:finetune:unsloth/mi... Base model:unsloth/mistral-sma... Bn Conversational De Decensored En Es Fa Fr Heretic Hi Id Image-text-to-text Instruct It Ja Ko Mistral3 Ms Ne Openrail++ Pl Pt Region:us Reproducible Ro Ru Safetensors Sharded Sr Sv Tensorflow Tr Uk Uncensored Vi Vllm Zh

Model Card on HF 🤗: https://huggingface.co/richardyoung/Mistral-Small-3.2-24B-Instruct-2506-ARA-heretic

Mistral Small 3.2 24B Instruct 2506 ARA Heretic Parameters and Internals

LLM Name	Mistral Small 3.2 24B Instruct 2506 ARA Heretic
Repository 🤗	https://huggingface.co/richardyoung/Mistral-Small-3.2-24B-Instruct-2506-ARA-heretic
Base Model(s)	unsloth/Mistral-Small-3.2-24B-Instruct-2506 unsloth/Mistral-Small-3.2-24B-Instruct-2506
Model Size	24b
Required VRAM	48.2 GB
Updated	2026-11-10
Maintainer	richardyoung
Model Type	mistral3
Instruction-Based	Yes
Model Files	4.9 GB: 1-of-10 4.8 GB: 2-of-10 4.8 GB: 3-of-10 4.9 GB: 4-of-10 4.8 GB: 5-of-10 4.8 GB: 6-of-10 4.9 GB: 7-of-10 4.8 GB: 8-of-10 4.8 GB: 9-of-10 4.7 GB: 10-of-10
Supported Languages	en fr de es pt it ja ko ru zh ar fa id ms ne pl ro sr sv tr uk vi hi bn
Model Architecture	Mistral3ForConditionalGeneration
License	apache-2.0
Model Max Length	131072
Transformers Version	5.18.0
Tokenizer Class	TokenizersBackend
Padding Token	<pad>

Rank the Mistral Small 3.2 24B Instruct 2506 ARA Heretic 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 Mistral Small 3.2 24B Instruct 2506 ARA Heretic

Best Alternatives	Context / RAM	Downloads	Likes
...resolvable Pattern Finetune V1	0K / 48.2 GB	55	0
... 2 24B Instruct 2512 MLX MXFP4	0K / 12.5 GB	1377	4
...all 3.2 24B Instruct 2506 Bf16	0K / 46.9 GB	128	1
...3.2 24B Instruct 2506 Bnb 4bit	0K / 14.5 GB	7430	10
...Instruct 2506 Unsloth Bnb 4bit	0K / 14.6 GB	2107	13
...all 3.2 24B Instruct 2506 8bit	0K / 25 GB	320	4
...2 24B Instruct 2512 OptiQ 4bit	0K / 16.5 GB	422	1
...l 2 24B Instruct 2512 MLX 8Bit	0K / 25 GB	3915	6
...all 3.2 24B Instruct 2506 4bit	0K / 13.3 GB	682	1
Devstral Small 2507 GGUF	0K / 0.9 GB	4818	80

Note: green Score (e.g. "73.2") means that the model is better than richardyoung/Mistral-Small-3.2-24B-Instruct-2506-ARA-heretic.