What are the hardware requirements for EleutherAI GPT Neox 20B 4bits?

EleutherAI GPT Neox 20B 4bits requires approximately 12.5 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 EleutherAI GPT Neox 20B 4bits and how large is it?

EleutherAI GPT Neox 20B 4bits is developed by RichardErkhov, a model with 20b 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 EleutherAI GPT Neox 20B 4bits?

EleutherAI GPT Neox 20B 4bits is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

EleutherAI GPT Neox 20B 4bits by RichardErkhov — VRAM 12.5GB, 2K context

Name: EleutherAI GPT Neox 20B 4bits
Author: RichardErkhov

EleutherAI GPT Neox 20B 4bits is an open-source language model by RichardErkhov. Features: 20b LLM, VRAM: 12.5GB, Context: 2K, License: apache-2.0, LLM Explorer Score: 0.12.

Arxiv:2101.00027 Arxiv:2104.09864 Arxiv:2201.07311 Arxiv:2204.06745 4-bit Bitsandbytes Endpoints compatible Gpt neox Pytorch Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/RichardErkhov/EleutherAI_-_gpt-neox-20b-4bits

EleutherAI GPT Neox 20B 4bits Parameters and Internals

Model Type

Transformer-based Language Model, causal-lm

Use Cases

Areas:

Research, Scientific uses

Primary Use Cases:

Extract features useful for downstream tasks

Limitations:

Not intended for deployment as-is, Not fine-tuned for downstream tasks such as writing genre prose or commercial chatbots, English-language only

Considerations:

Use in accordance with the Apache 2.0 license. Conduct your own risk and bias assessment when fine-tuning for specific tasks.

Additional Notes

The model was trained on 'the Pile', which was not deduplicated before use. Contains texts with biases regarding gender, religion, and race. Curating outputs before presenting to a human reader is recommended.

Supported Languages

en (Full)

Training Details

Data Sources:

EleutherAI/pile

Methodology:

Autoregressive training using GPT-NeoX library

Context Length:

2048

Model Architecture:

Architecture resembles GPT-3 and is almost identical to that of GPT-J-6B.

LLM Name	EleutherAI GPT Neox 20B 4bits
Repository 🤗	https://huggingface.co/RichardErkhov/EleutherAI_-_gpt-neox-20b-4bits
Model Size	20b
Required VRAM	12.5 GB
Updated	2026-01-03
Maintainer	RichardErkhov
Model Type	gpt_neox
Model Files	5.0 GB: 1-of-3 4.9 GB: 2-of-3 2.6 GB: 3-of-3
Supported Languages	en
Model Architecture	GPTNeoXForCausalLM
License	apache-2.0
Context Length	2048
Model Max Length	2048
Transformers Version	4.39.3
Tokenizer Class	GPTNeoXTokenizer
Vocabulary Size	50432
Torch Data Type	float16

Rank the EleutherAI GPT Neox 20B 4bits 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 EleutherAI GPT Neox 20B 4bits

Best Alternatives	Context / RAM	Downloads	Likes
GPT Neox 20B	2K / 40.8 GB	794665	583
GPT NeoXT Chat Base 20B	2K / 41.2 GB	256	694
H2ogpt Oasst1 512 20B	2K / 41.2 GB	1064	39
...t Gm Oasst1 Multilang 1024 20B	2K / 41.2 GB	140	10
H2ogpt Gm Oasst1 En 1024 20B	2K / 41.2 GB	55	4
GPT Neox 20B Full Precision	2K / 82.5 GB	52	0
Oasst GPT Neox 20B 1000 Steps	2K / 41.2 GB	75	0
Oasst GPT Neox 20B 3000 Steps	2K / 41.2 GB	60	0
GPTNeoX 20B TestGen Dart V1.0	2K / 41.2 GB	5	2
GPT NeoX 20B Erebus	2K / 41.4 GB	576	88

Note: green Score (e.g. "73.2") means that the model is better than RichardErkhov/EleutherAI_-_gpt-neox-20b-4bits.