What are the hardware requirements for Llama2 22B GPLATTY GGML?

Llama2 22B GPLATTY GGML requires approximately 9.2 GB of VRAM. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Llama2 22B GPLATTY GGML and how large is it?

Llama2 22B GPLATTY GGML is developed by TheBloke, a model with 22b 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 Llama2 22B GPLATTY GGML?

Llama2 22B GPLATTY GGML is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Llama2 22B GPLATTY GGML by TheBloke — VRAM 9.2GB

Name: Llama2 22B GPLATTY GGML
Author: TheBloke

Llama2 22B GPLATTY GGML is an open-source language model by TheBloke. Features: 22b LLM, VRAM: 9.2GB, License: llama2, Quantized, LLM Explorer Score: 0.08.

Base model:finetune:grimpep/ll... Base model:grimpep/llama2-22b-... Ggml Llama Llama2 Quantized Region:us

Model Card on HF 🤗: https://huggingface.co/TheBloke/Llama2-22B-GPLATTY-GGML

Llama2 22B GPLATTY GGML Parameters and Internals

Model Type

llama

Additional Notes

The model is experimental, created by merging different Llama versions with a donor model. GGML formatted model files are used for CPU+GPU inference.

Training Details

Methodology:

Merged using Frankenstein script.

Hardware Used:

160GB of system RAM for model merging

Release Notes

Version:

GGML format

Notes:

Deprecated and to be replaced by GGUF format as of August 21st, 2023.

Version:

GGUF format support

Notes:

Suggested for compatibility with latest llama.cpp.

LLM Name	Llama2 22B GPLATTY GGML
Repository 🤗	https://huggingface.co/TheBloke/Llama2-22B-GPLATTY-GGML
Model Name	Llama2 22B GPLATTY
Model Creator	grimpep
Base Model(s)	grimpep/llama2-22B-GPLATTY grimpep/llama2-22B-GPLATTY
Model Size	22b
Required VRAM	9.2 GB
Updated	2026-08-04
Maintainer	TheBloke
Model Type	llama
Model Files	9.2 GB 11.6 GB 10.6 GB 9.5 GB 12.3 GB 13.7 GB 13.2 GB 12.3 GB 15.0 GB 16.4 GB 15.5 GB 15.0 GB 17.9 GB 23.1 GB
GGML Quantization	Yes
Quantization Type	ggml
Model Architecture	AutoModel
License	llama2

Rank the Llama2 22B GPLATTY GGML 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 Llama2 22B GPLATTY GGML

Best Alternatives	Context / RAM	Downloads	Likes
Huginn 22B Prototype GGML	0K / 9.2 GB	4	1
Llama2 22B Daydreamer V3 GGML	0K / 9.2 GB	5	6
Llama2 22B Daydreamer V2 GGML	0K / 9.2 GB	5	4
Codestral 22B V0.1 GGUF	0K / 4.8 GB	2720	9
...t2cypher Codestral Q4 K M Gguf	0K / 13.3 GB	160	5
...xt2cypher Codestral 16bit Gguf	0K / 44.5 GB	47	0
Qwen1.5 22B Chat Merge GGUF	0K / 12.6 GB	32	1
Nucleus 22B Token 500B GGUF	0K / 9.1 GB	345	3
Llama2 22B Daydreamer V3 GGUF	0K / 9.1 GB	362	2
Llama2 22B GPLATTY GGUF	0K / 9.1 GB	168	3

Note: green Score (e.g. "73.2") means that the model is better than TheBloke/Llama2-22B-GPLATTY-GGML.