What are the hardware requirements for Nous Hermes Llama2 70B?

Nous Hermes Llama2 70B requires approximately 138 GB of VRAM. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Nous Hermes Llama2 70B and how large is it?

Nous Hermes Llama2 70B is developed by NousResearch, a model with 70b 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 Nous Hermes Llama2 70B?

Nous Hermes Llama2 70B is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Nous Hermes Llama2 70B by NousResearch — VRAM 138GB

Name: Nous Hermes Llama2 70B
Author: NousResearch

Nous Hermes Llama2 70B is an open-source language model by NousResearch. Features: 70b LLM, VRAM: 138GB, License: mit, LLM Explorer Score: 0.09.

Distillation En Endpoints compatible Llama Llama2 Lora Pytorch Region:us Self-instruct Sharded Synthetic instruction

Model Card on HF 🤗: https://huggingface.co/NousResearch/Nous-Hermes-Llama2-70b

Nous Hermes Llama2 70B Parameters and Internals

Model Type

language model

Additional Notes

The model is known for its long responses and low hallucination rates. GPU training involved an 8x H100 80GB machine.

Training Details

Data Sources:

GPTeacher, Nous Instruct Dataset, GPT4-LLM, Unnatural Instructions, Airoboros dataset, Camel-AI's domain expert datasets, CodeAlpaca dataset

Methodology:

Fine-tuned on synthetic GPT-4 outputs on an 8x H100 80GB machine

Context Length:

4096

Input Output

Input Format:

### Instruction: ### Input: ### Response:

LLM Name	Nous Hermes Llama2 70B
Repository 🤗	https://huggingface.co/NousResearch/Nous-Hermes-Llama2-70b
Model Size	70b
Required VRAM	138 GB
Updated	2026-08-04
Maintainer	NousResearch
Model Files	1.7 GB 9.8 GB: 1-of-15 9.8 GB: 2-of-15 10.0 GB: 3-of-15 9.8 GB: 4-of-15 9.8 GB: 5-of-15 9.8 GB: 6-of-15 10.0 GB: 7-of-15 9.8 GB: 8-of-15 9.8 GB: 9-of-15 9.8 GB: 10-of-15 10.0 GB: 11-of-15 9.8 GB: 12-of-15 9.8 GB: 13-of-15 9.5 GB: 14-of-15 0.5 GB: 15-of-15
Supported Languages	en
Model Architecture	AutoModelForCausalLM
License	mit
Is Biased	none
Tokenizer Class	LlamaTokenizer
Beginning of Sentence Token	<s>
End of Sentence Token	</s>
Unk Token	<unk>
PEFT Type	LORA
LoRA Model	Yes
PEFT Target Modules	o_proj\|q_proj\|up_proj\|k_proj\|gate_proj\|down_proj\|v_proj
LoRA Alpha	64
LoRA Dropout	0.1
R Param	32

Rank the Nous Hermes Llama2 70B 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 Llama2 70B

Model	Likes	Downloads	VRAM
Nous Hermes Llama2 70B GGUF	27	836	29 GB
Nous Hermes Llama2 70B AWQ	2	6	36 GB
Nous Hermes Llama2 70B GGML	13	4	29 GB
Nous Hermes Llama2 70B GPTQ	7	8	35 GB

Best Alternatives to Nous Hermes Llama2 70B

Best Alternatives	Context / RAM	Downloads	Likes
Llama 3.1 Tango 70B	0K / 3 GB	6	8
LLama3 70B SWE LLM	0K / 16.1 GB	3	1
Limarpv3 Llama2 70B Qlora	0K / 1.7 GB	1811	2
Limarpv3 Llama2 70B Qlora	0K / 1.7 GB	16	2
Llama 2 70B Chat Longlora 32K	0K / 0.1 GB	6	9
Llama 2 70B Longlora 32K	0K / 0.1 GB	6	18
... 70M Instruct Orca Chkpt 64000	0K / 0.2 GB	8	1
NorskGPT Llama 3 70B Adapter	0K / 0.2 GB	6	7
Llama 3 70B Tagengo	0K / 141.9 GB	3	1
...3 70B Instruct Uncensored Lora	0K / 0.8 GB	8	3

Note: green Score (e.g. "73.2") means that the model is better than NousResearch/Nous-Hermes-Llama2-70b.