What are the hardware requirements for Nucleus 22B Token 500B GPTQ?

Nucleus 22B Token 500B GPTQ requires approximately 11.6 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 Nucleus 22B Token 500B GPTQ and how large is it?

Nucleus 22B Token 500B GPTQ 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 Nucleus 22B Token 500B GPTQ?

Nucleus 22B Token 500B GPTQ is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Nucleus 22B Token 500B GPTQ by TheBloke — VRAM 11.6GB, 2K context

Name: Nucleus 22B Token 500B GPTQ
Author: TheBloke

Nucleus 22B Token 500B GPTQ is an open-source language model by TheBloke. Features: 22b LLM, VRAM: 11.6GB, Context: 2K, License: mit, Quantized, LLM Explorer Score: 0.1.

4-bit Base model:nucleusai/nucleus-2... Base model:quantized:nucleusai... En Gptq Llama Quantized Region:us Safetensors

Model Card on HF 🤗: https://huggingface.co/TheBloke/nucleus-22B-token-500B-GPTQ

Nucleus 22B Token 500B GPTQ Parameters and Internals

Model Type

causal decoder-only

Use Cases

Areas:

Research

Applications:

Large language models, Summarization, Text generation, Chatbots

Primary Use Cases:

Foundation for further specialization and finetuning

Limitations:

Not suitable for production use without adequate assessment of risks

Considerations:

Bias and stereotypical content may exist.

Additional Notes

The model performs well compared to similar-sized open-source models owing to its training on a large corpus.

Supported Languages

English (primary)

Training Details

Data Sources:

RefinedWeb-English, Books, Code, Technical, Math

Data Volume:

500B tokens

Context Length:

2048

Training Time:

two weeks in early August 2023

Hardware Used:

256 A100 80GB GPUs

Model Architecture:

similar to Llama [7B]

Responsible Ai Considerations

Fairness:

The model may carry stereotypes and biases from the web.

Mitigation Strategies:

Finetuning for specific tasks and implementing guardrails for production use are recommended.

LLM Name	Nucleus 22B Token 500B GPTQ
Repository 🤗	https://huggingface.co/TheBloke/nucleus-22B-token-500B-GPTQ
Model Name	Nucleus 22B Token 500B
Model Creator	NucleusAI
Base Model(s)	Nucleus 22B Token 500B NucleusAI/nucleus-22B-token-500B
Model Size	22b
Required VRAM	11.6 GB
Updated	2026-07-02
Maintainer	TheBloke
Model Type	llama
Model Files	11.6 GB
Supported Languages	en
GPTQ Quantization	Yes
Quantization Type	gptq
Model Architecture	LlamaForCausalLM
License	mit
Context Length	2048
Model Max Length	2048
Transformers Version	4.35.2
Tokenizer Class	LlamaTokenizer
Beginning of Sentence Token	<s>
End of Sentence Token	</s>
Unk Token	<unk>
Vocabulary Size	32000
Torch Data Type	float16

Rank the Nucleus 22B Token 500B GPTQ 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 Nucleus 22B Token 500B GPTQ

Best Alternatives	Context / RAM	Downloads	Likes
Llama2 22B Daydreamer V3 GPT	4K / 12 GB	7	0
Llama2 22B Daydreamer V3 GPTQ	4K / 12 GB	10	6
Llama2 22B Daydreamer V2 GPTQ	4K / 12 GB	8	8
Ypotryll 22B Gptq	4K / 12 GB	9	1
Llama2 22B GPLATTY GPT	4K / 12 GB	6	0
Huginn 22B Prototype GPTQ	4K / 12 GB	2	1
Llama2 22B GPLATTY GPTQ	4K / 12 GB	4	4
...LLM 22B Instruct 2512 Mlx 4bit	32K / 12.7 GB	127	3
...LLM 22B Instruct 2512 Mlx 8bit	32K / 24.1 GB	53	1
Dendrite 22Bchk2 F16	4K / 43.7 GB	7	1

Note: green Score (e.g. "73.2") means that the model is better than TheBloke/nucleus-22B-token-500B-GPTQ.