What are the hardware requirements for Nous Hermes 2 Mixtral 8x7B SFT?

Nous Hermes 2 Mixtral 8x7B SFT requires approximately 93.6 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 Nous Hermes 2 Mixtral 8x7B SFT and how large is it?

Nous Hermes 2 Mixtral 8x7B SFT is developed by NousResearch, a model with 46.7b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

How does Nous Hermes 2 Mixtral 8x7B SFT perform on standard benchmarks?

Nous Hermes 2 Mixtral 8x7B SFT has the following published scores: MMLU 72.21. Compare against reference models on this page or on the LLM Explorer leaderboards.

Nous Hermes 2 Mixtral 8x7B SFT by NousResearch — VRAM 93.6GB, 32K context

Name: Nous Hermes 2 Mixtral 8x7B SFT
Author: NousResearch

Nous Hermes 2 Mixtral 8x7B SFT is an open-source language model by NousResearch. Features: 46.7b LLM, VRAM: 93.6GB, Context: 32K, License: apache-2.0, MoE, LLM Explorer Score: 0.18, Arc: 69.7, HellaSwag: 86.7, MMLU: 72.2, GSM8K: 69.6.

Base model:finetune:mistralai/... Base model:mistralai/mixtral-8... Chatml Conversational Dataset:teknium/openhermes-2.5 Deploy:azure Distillation En Endpoints compatible Finetuned Gpt4 Instruct Mixtral Moe Region:us Safetensors Sharded Synthetic data Tensorflow

Model Card on HF 🤗: https://huggingface.co/NousResearch/Nous-Hermes-2-Mixtral-8x7B-SFT

Nous Hermes 2 Mixtral 8x7B SFT Parameters and Internals

Model Type

instruct, finetune, chatml, gpt4, synthetic data, distillation

Use Cases

Areas:

research, commercial applications

Applications:

finetuning and demonstrations

Primary Use Cases:

chat-oriented tasks

Limitations:

Requires more than 24GB of VRAM for inference

Considerations:

Use quantized models for lower VRAM devices.

Additional Notes

First SFT version; Quantized GGUF models recommended for better performance with lower VRAM.

Supported Languages

en (high)

Training Details

Data Sources:

GPT-4 generated data, open datasets

Data Volume:

1,000,000 entries

Methodology:

Supervised finetune only

Hardware Used:

Sponsored by Together.ai

Model Architecture:

Mixtral 8x7B MoE LLM

Input Output

Input Format:

ChatML; system prompts allow steerability.

Accepted Modalities:

text

Output Format:

text

Performance Tips:

Use quantized versions for reduced VRAM requirements.

LLM Name	Nous Hermes 2 Mixtral 8x7B SFT
Repository 🤗	https://huggingface.co/NousResearch/Nous-Hermes-2-Mixtral-8x7B-SFT
Base Model(s)	mistralai/Mixtral-8x7B-v0.1 mistralai/Mixtral-8x7B-v0.1
Model Size	46.7b
Required VRAM	93.6 GB
Updated	2026-07-30
Maintainer	NousResearch
Model Type	mixtral
Model Files	4.9 GB: 1-of-19 5.0 GB: 2-of-19 5.0 GB: 3-of-19 4.9 GB: 4-of-19 5.0 GB: 5-of-19 5.0 GB: 6-of-19 4.9 GB: 7-of-19 5.0 GB: 8-of-19 5.0 GB: 9-of-19 4.9 GB: 10-of-19 5.0 GB: 11-of-19 5.0 GB: 12-of-19 5.0 GB: 13-of-19 4.9 GB: 14-of-19 5.0 GB: 15-of-19 5.0 GB: 16-of-19 4.9 GB: 17-of-19 5.0 GB: 18-of-19 4.2 GB: 19-of-19
Supported Languages	en
Model Architecture	MixtralForCausalLM
License	apache-2.0
Context Length	32768
Model Max Length	32768
Transformers Version	4.36.0.dev0
Tokenizer Class	LlamaTokenizer
Padding Token	</s>
Vocabulary Size	32002
Torch Data Type	bfloat16

Rank the Nous Hermes 2 Mixtral 8x7B SFT 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

Quantized Models of the Nous Hermes 2 Mixtral 8x7B SFT

Model	Likes	Downloads	VRAM
...Hermes 2 Mixtral 8x7B SFT GGUF	4	788	17 GB
...Hermes 2 Mixtral 8x7B SFT GGUF	25	498	17 GB
...Hermes 2 Mixtral 8x7B SFT GPTQ	11	10	23 GB
... Hermes 2 Mixtral 8x7B SFT AWQ	3	6	24 GB

Best Alternatives to Nous Hermes 2 Mixtral 8x7B SFT

Best Alternatives	Context / RAM	Downloads	Likes
Mixtral 8x7B Instruct V0.1	32K / 93.6 GB	842893	4720
Nous Hermes 2 Mixtral 8x7B DPO	32K / 93.6 GB	40865	457
Mixtral 8x7B V0.1	32K / 93.6 GB	49231	1820
Sensualize Mixtral Bf16	32K / 93.6 GB	0	0
Skadi Mixtral V1	32K / 93.5 GB	0	0
Franziska Mixtral V1	32K / 93.5 GB	0	0
Typhon Mixtral V1	32K / 93.4 GB	0	0
GritLM 8x7B KTO	32K / 93.6 GB	7976	3
Mixtral 8x7B Instruct V0.1	32K / 93.6 GB	7	0
Smaug Mixtral V0.1	32K / 187.7 GB	8437	12

Note: green Score (e.g. "73.2") means that the model is better than NousResearch/Nous-Hermes-2-Mixtral-8x7B-SFT.