What are the hardware requirements for Llama 65B Instruct?

Llama 65B Instruct requires approximately 130.4 GB of VRAM and supports a context window of 2K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Llama 65B Instruct and how large is it?

Llama 65B Instruct is developed by upstage, a model with 65b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

How does Llama 65B Instruct perform on standard benchmarks?

Llama 65B Instruct has the following published scores: MMLU 64.77. Compare against reference models on this page or on the LLM Explorer leaderboards.

Llama 65B Instruct by upstage — VRAM 130.4GB, 2K context

Name: Llama 65B Instruct
Author: upstage

Llama 65B Instruct is an open-source language model by upstage. Features: 65b LLM, VRAM: 130.4GB, Context: 2K, Instruction-Based, LLM Explorer Score: 0.09, Arc: 68.9, HellaSwag: 86.4, MMLU: 64.8, GSM8K: 26.2.

Deploy:azure En Endpoints compatible Instruct Instruction Llama Pytorch Region:us Sharded Upstage

Model Card on HF 🤗: https://huggingface.co/upstage/llama-65b-instruct

Llama 65B Instruct Parameters and Internals

Model Type

text generation

Use Cases

Areas:

research, commercial applications

Limitations:

Known limitations or unsuitable use cases not provided.

Additional Notes

Tested for handling up to 10k+ input tokens.

Supported Languages

English (proficient)

Training Details

Data Sources:

Orca-style dataset

Methodology:

fine-tuning

Hardware Used:

A100x8 * 4

Input Output

Input Format:

Prompt format includes System, User, and Assistant text.

Accepted Modalities:

text

Output Format:

text generation output

Performance Tips:

Use 'rope_scaling' option for handling longer inputs.

LLM Name	Llama 65B Instruct
Repository 🤗	https://huggingface.co/upstage/llama-65b-instruct
Model Size	65b
Required VRAM	130.4 GB
Updated	2026-07-18
Maintainer	upstage
Model Type	llama
Instruction-Based	Yes
Model Files	9.9 GB: 1-of-14 9.7 GB: 2-of-14 9.7 GB: 3-of-14 9.7 GB: 4-of-14 9.7 GB: 5-of-14 9.7 GB: 6-of-14 9.7 GB: 7-of-14 9.7 GB: 8-of-14 9.7 GB: 9-of-14 9.7 GB: 10-of-14 9.7 GB: 11-of-14 9.7 GB: 12-of-14 9.7 GB: 13-of-14 4.1 GB: 14-of-14
Supported Languages	en
Model Architecture	LlamaForCausalLM
Context Length	2048
Model Max Length	2048
Transformers Version	4.30.2
Tokenizer Class	LlamaTokenizer
Beginning of Sentence Token	<s>
End of Sentence Token	</s>
Unk Token	<unk>
Vocabulary Size	32000
Torch Data Type	bfloat16

Rank the Llama 65B Instruct 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 Llama 65B Instruct

Model	Likes	Downloads	VRAM
...stage Llama1 65B Instruct GGUF	1	1110	27 GB
...pstage Llama1 65B Instruct AWQ	1	9	34 GB
...stage Llama1 65B Instruct GPTQ	3	9	34 GB

Best Alternatives to Llama 65B Instruct

Best Alternatives	Context / RAM	Downloads	Likes
Lite Oute 1 65M Instruct	2K / 0.3 GB	15	12
Open Instruct Human Mix 65B	2K / 260.8 GB	173	4
Open Instruct Sharegpt 65B	2K / 260.8 GB	45	2
...pen Instruct Human Mix 65B AWQ	2K / 34.7 GB	13	0
...en Instruct Human Mix 65B GPTQ	2K / 33.5 GB	11	0
...pstage Llama1 65B Instruct AWQ	2K / 34.7 GB	9	1
...stage Llama1 65B Instruct GPTQ	2K / 34.7 GB	9	3

Note: green Score (e.g. "73.2") means that the model is better than upstage/llama-65b-instruct.