What are the hardware requirements for MicroLlama?

MicroLlama requires approximately 1.2 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 MicroLlama and how large is it?

MicroLlama is developed by keeeeenw, a model with 627b 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 MicroLlama?

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

MicroLlama by keeeeenw — VRAM 1.2GB, 2K context

Name: MicroLlama
Author: keeeeenw

MicroLlama is an open-source language model by keeeeenw. Features: 627b LLM, VRAM: 1.2GB, Context: 2K, License: apache-2.0, LLM Explorer Score: 0.14.

Arxiv:2401.02385 Dataset:cerebras/slimpajama-62... Deploy:azure En Endpoints compatible Llama Model-index Region:us Safetensors

Model Card on HF 🤗: https://huggingface.co/keeeeenw/MicroLlama

MicroLlama Benchmarks

LLME Score: 0.14336

^nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").

What is the LLM Explorer Rank (Score)

🌟 Advertise your project 🚀

MicroLlama Parameters and Internals

Model Type

text-generation

Additional Notes

Project evolved from TinyLlama; a smaller model setup was developed to fit budget and resource constraints.

Supported Languages

en (NLP)

Training Details

Data Sources:

cerebras/SlimPajama-627B

Data Volume:

50B tokens

Context Length:

2048

Hardware Used:

4 x Nvidia 4090

Model Architecture:

block_size=2048, vocab_size=32000, padding_multiple=64, n_layer=12, n_head=16, n_embd=1024, rotary_percentage=1.0, parallel_residual=False, bias=False, _norm_class="FusedRMSNorm", norm_eps=1e-5, _mlp_class="LLaMAMLP", intermediate_size=5632, n_query_groups=4

LLM Name	MicroLlama
Repository 🤗	https://huggingface.co/keeeeenw/MicroLlama
Model Size	627b
Required VRAM	1.2 GB
Updated	2026-04-23
Maintainer	keeeeenw
Model Type	llama
Model Files	1.2 GB
Supported Languages	en
Model Architecture	LlamaForCausalLM
License	apache-2.0
Context Length	2048
Model Max Length	2048
Transformers Version	4.39.1
Tokenizer Class	LlamaTokenizer
Vocabulary Size	32000
Torch Data Type	float32

Quantized Models of the MicroLlama

Model	Likes	Downloads	VRAM
...llama Python Instruct 0.3 4bit	0	5	0 GB

Best Alternatives to MicroLlama

Best Alternatives	Context / RAM	Downloads	Likes
TinyLlama V1.1	2K / 4.4 GB	12664	111
CroissantLLMBase	2K / 5.4 GB	1460	34
...Llama 1.1B 1.5T OpenOrca Alpha	2K / 2.2 GB	917	4
TinyLlama 1.1B 1T OpenOrca	2K / 2.2 GB	912	7
TinyLlama V1.1 Math Code	2K / 4.4 GB	1288	14
TinyLlama V1.1 Chinese	2K / 4.4 GB	881	14
TinyLlama 1.1B 1T OpenOrca AWQ	2K / 0.8 GB	17	2
...inyLlama 1.1B 1T OpenOrca GPTQ	2K / 0.8 GB	9	2

Note: green Score (e.g. "73.2") means that the model is better than keeeeenw/MicroLlama.

Rank the MicroLlama Capabilities

🆘 Have you tried this model? Rate its performance. This feedback would greatly assist ML community in identifying the most suitable model for their needs. Your contribution really does make a difference! 🌟

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

What open-source LLMs or SLMs are you in search of? 53999 in total.

Email us: info@extractum.io. Our Privacy Policy | Terms and Conditions | Suggest an improvement.

Our Social Media →

Original data from HuggingFace, OpenCompass and various public git repos.

Check out Ag3ntum — our secure, self-hosted AI agent for server management.

Release v20260328a

Support LLM Explorer

MicroLlama by keeeeenw