What are the hardware requirements for Mistral Large 4 1T A52B FP8?

Mistral Large 4 1T A52B FP8 requires approximately 880 GB of VRAM. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Mistral Large 4 1T A52B FP8 and how large is it?

Mistral Large 4 1T A52B FP8 is developed by unsloth. 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 Large 4 1T A52B FP8?

Mistral Large 4 1T A52B FP8 is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Mistral Large 4 1T A52B FP8 by unsloth — VRAM 880GB

Name: Mistral Large 4 1T A52B FP8
Author: unsloth

Mistral Large 4 1T A52B FP8 is an open-source language model by unsloth. Features: LLM, VRAM: 880GB, License: other, LLM Explorer Score: 0.33.

Ar Base model:mistralai/mistral-l... Base model:quantized:mistralai... Bg Bn Conversational Cs Da De El En Endpoints compatible Es Et Fa Fi Fp8 Fr Ga Hi Hr Hu Id Image-text-to-text It Ja Ko Llama.cpp Lt Lv Mistral Mistral large4 Ms Mt Ne Nl Pl Pt Region:us Ro Ru Sglang Sharded Sk Sl Sr Sv Tensorflow Tr Uk Unsloth Vi Vllm Zh

Model Card on HF 🤗: https://huggingface.co/unsloth/Mistral-Large-4-1T-A52B-FP8

Mistral Large 4 1T A52B FP8 Parameters and Internals

LLM Name	Mistral Large 4 1T A52B FP8
Repository 🤗	https://huggingface.co/unsloth/Mistral-Large-4-1T-A52B-FP8
Base Model(s)	Mistral Large 4 1T A52B mistralai/Mistral-Large-4-1T-A52B
Required VRAM	880 GB
Updated	2026-10-31
Maintainer	unsloth
Model Type	mistral_large4
Model Files	20.0 GB: 1-of-54 20.0 GB: 2-of-54 20.0 GB: 3-of-54 20.0 GB: 4-of-54 20.0 GB: 5-of-54 20.0 GB: 6-of-54 20.0 GB: 7-of-54 20.0 GB: 8-of-54 20.0 GB: 9-of-54 20.0 GB: 10-of-54 20.0 GB: 11-of-54 20.0 GB: 12-of-54 20.0 GB: 13-of-54 20.0 GB: 14-of-54 20.0 GB: 15-of-54 20.0 GB: 16-of-54 20.0 GB: 17-of-54 20.0 GB: 18-of-54 20.0 GB: 19-of-54 20.0 GB: 20-of-54 20.0 GB: 21-of-54 20.0 GB: 22-of-54 20.0 GB: 23-of-54 20.0 GB: 24-of-54 20.0 GB: 25-of-54 20.0 GB: 26-of-54 20.0 GB: 27-of-54 20.0 GB: 28-of-54 20.0 GB: 29-of-54 20.0 GB: 30-of-54 20.0 GB: 31-of-54 20.0 GB: 32-of-54 20.0 GB: 33-of-54 20.0 GB: 34-of-54 20.0 GB: 35-of-54 20.0 GB: 36-of-54 20.0 GB: 37-of-54 20.0 GB: 38-of-54 20.0 GB: 39-of-54 20.0 GB: 40-of-54 20.0 GB: 41-of-54 20.0 GB: 42-of-54 20.0 GB: 43-of-54 20.0 GB: 44-of-54
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 bg hr cs da nl et fi el hu ga lv lt mt sk sl
Model Architecture	MistralLarge4ForConditionalGeneration
License	other
Transformers Version	5.21.0.dev0

Rank the Mistral Large 4 1T A52B FP8 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 Large 4 1T A52B FP8

Best Alternatives	Context / RAM	Downloads	Likes
Mistral Large 4 1T A52B	0K / 880 GB	2945	1398

Note: green Score (e.g. "73.2") means that the model is better than unsloth/Mistral-Large-4-1T-A52B-FP8.