What are the hardware requirements for Bigstral 12B 32K 8xMoE?

Bigstral 12B 32K 8xMoE requires approximately 163.3 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 Bigstral 12B 32K 8xMoE and how large is it?

Bigstral 12B 32K 8xMoE is developed by bartowski, 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 Bigstral 12B 32K 8xMoE?

Bigstral 12B 32K 8xMoE is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Bigstral 12B 32K 8xMoE by bartowski — VRAM 163.3GB, 32K context

Name: Bigstral 12B 32K 8xMoE
Author: bartowski

Bigstral 12B 32K 8xMoE is an open-source language model by bartowski. Features: 7b LLM, VRAM: 163.3GB, Context: 32K, Instruction-Based, LLM Explorer Score: 0.11.

Base model:finetune:mistralai/... Base model:mistralai/mistral-7... Conversational Endpoints compatible Instruct Merge Mergekit Mixtral Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/bartowski/bigstral-12b-32k-8xMoE

Bigstral 12B 32K 8xMoE Parameters and Internals

Additional Notes

The model is created using Mergekit MoE branch with abacusai/bigstral-12b-32k as a base model. It uses MoE configuration with random gate_mode, and incorporates numerous expert models from the same source without specific prompt modifications.

LLM Name	Bigstral 12B 32K 8xMoE
Repository 🤗	https://huggingface.co/bartowski/bigstral-12b-32k-8xMoE
Base Model(s)	mistralai/Mistral-7B-Instruct-v0.2 mistralai/Mistral-7B-Instruct-v0.2
Model Size	7b
Required VRAM	163.3 GB
Updated	2026-05-10
Maintainer	bartowski
Model Type	mixtral
Instruction-Based	Yes
Model Files	9.9 GB: 1-of-17 10.0 GB: 2-of-17 10.0 GB: 3-of-17 10.0 GB: 4-of-17 10.0 GB: 5-of-17 10.0 GB: 6-of-17 10.0 GB: 7-of-17 10.0 GB: 8-of-17 10.0 GB: 9-of-17 10.0 GB: 10-of-17 10.0 GB: 11-of-17 10.0 GB: 12-of-17 10.0 GB: 13-of-17 10.0 GB: 14-of-17 10.0 GB: 15-of-17 10.0 GB: 16-of-17 3.4 GB: 17-of-17
Model Architecture	MixtralForCausalLM
Context Length	32768
Model Max Length	32768
Transformers Version	4.38.2
Tokenizer Class	LlamaTokenizer
Padding Token	<s>
Vocabulary Size	32000
Torch Data Type	bfloat16

Rank the Bigstral 12B 32K 8xMoE 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 Bigstral 12B 32K 8xMoE

Best Alternatives	Context / RAM	Downloads	Likes
Mini Mixtral V0.2	32K / 25.8 GB	3923	4
OpenMistral MoE	32K / 48.3 GB	1218	0
Multilingual Mistral	32K / 93.5 GB	65	2
Magiq 3	32K / 37.1 GB	5	3
FNCARL9000	32K / 48.3 GB	8	0
CollAIborate4x7B	32K / 48.7 GB	6	1
CollAIborate4x7B	32K / 48.7 GB	1	1
OpenMistral MoE	32K / 48.3 GB	6	0
...t V0.2 2x7B MoE 6.0bpw H6 EXL2	32K / 9.9 GB	3	1

Note: green Score (e.g. "73.2") means that the model is better than bartowski/bigstral-12b-32k-8xMoE.