What are the hardware requirements for Hiraya 1 H MoE?

Hiraya 1 H MoE requires approximately 3.6 GB of VRAM. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Hiraya 1 H MoE and how large is it?

Hiraya 1 H MoE is developed by DeepNarra, a model with 908.6m 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 Hiraya 1 H MoE?

Hiraya 1 H MoE is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Hiraya 1 H MoE by DeepNarra — VRAM 3.6GB

Name: Hiraya 1 H MoE
Author: DeepNarra

Hiraya 1 H MoE is an open-source language model by DeepNarra. Features: 908.6m LLM, VRAM: 3.6GB, License: apache-2.0, MoE, Instruction-Based, LLM Explorer Score: 0.31.

0.9b 908m 92-experts Base model:deepnarra/hiraya-1-... Base model:finetune:deepnarra/... Dataset:cohereforai/aya collec... Dataset:huggingfacefw/fineweb-... Dataset:m-a-p/codefeedback-fil... Deepnarra En Endpoints compatible H-moe Hiraya Instruct Moe Multilingual Region:us Safetensors V1.0 Zamba2

Model Card on HF 🤗: https://huggingface.co/DeepNarra/Hiraya-1-H-MoE

Hiraya 1 H MoE Parameters and Internals

LLM Name	Hiraya 1 H MoE
Repository 🤗	https://huggingface.co/DeepNarra/Hiraya-1-H-MoE
Model Name	H-MoE-Omni-Zamba2-Ascended
Base Model(s)	Hiraya 1 Base DeepNarra/Hiraya-1-Base
Model Size	908.6m
Required VRAM	3.6 GB
Updated	2026-10-28
Maintainer	DeepNarra
Instruction-Based	Yes
Model Files	3.6 GB
Supported Languages	en
Model Architecture	AutoModel
License	apache-2.0
Tokenizer Class	TokenizersBackend
Padding Token	<pad>
Vocabulary Size	48760

Rank the Hiraya 1 H MoE 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 Hiraya 1 H MoE

Best Alternatives	Context / RAM	Downloads	Likes
Hiraya 1 Instruct	0K / 3.5 GB	1220	0
Hiraya 1 Reasoning	0K / 1.8 GB	931	0

Note: green Score (e.g. "73.2") means that the model is better than DeepNarra/Hiraya-1-H-MoE.