What are the hardware requirements for ZAYA1 Base?

ZAYA1 Base requires approximately 17.7 GB of VRAM and supports a context window of 32K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed ZAYA1 Base and how large is it?

ZAYA1 Base is developed by Zyphra, a model with 8.8b 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 ZAYA1 Base?

ZAYA1 Base is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

ZAYA1 Base by Zyphra — VRAM 17.7GB, 32K context

Name: ZAYA1 Base
Author: Zyphra

ZAYA1 Base is an open-source language model by Zyphra. Features: 8.8b LLM, VRAM: 17.7GB, Context: 32K, License: apache-2.0.

Arxiv:2510.04476 Arxiv:2511.17127 Conversational Endpoints compatible Region:us Safetensors Sharded Tensorflow Zaya

Model Card on HF 🤗: https://huggingface.co/Zyphra/ZAYA1-base

ZAYA1 Base Parameters and Internals

LLM Name	ZAYA1 Base
Repository 🤗	https://huggingface.co/Zyphra/ZAYA1-base
Model Size	8.8b
Required VRAM	17.7 GB
Updated	2026-08-17
Maintainer	Zyphra
Model Type	zaya
Model Files	5.0 GB: 1-of-4 5.0 GB: 2-of-4 5.0 GB: 3-of-4 2.7 GB: 4-of-4
Model Architecture	ZayaForCausalLM
License	apache-2.0
Context Length	32768
Model Max Length	32768
Transformers Version	4.57.1
Tokenizer Class	GemmaTokenizerFast
Padding Token	<pad>
Vocabulary Size	262272
Torch Data Type	bfloat16

Rank the ZAYA1 Base 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