What are the hardware requirements for Llama 68M?

Llama 68M requires a context window of 2K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Llama 68M and how large is it?

Llama 68M is developed by Xenova, a model with 68m 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 Llama 68M?

Llama 68M is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Llama 68M by Xenova — 2K context

Name: Llama 68M
Author: Xenova

Llama 68M is an open-source language model by Xenova. Features: 68m LLM, Context: 2K, LLM Explorer Score: 0.09.

Base model:jackfram/llama-68m Base model:quantized:jackfram/... Llama Onnx Region:us Transformers.js

Model Card on HF 🤗: https://huggingface.co/Xenova/llama-68m

Llama 68M Parameters and Internals

Additional Notes

The model uses ONNX weights to be compatible with Transformers.js. Having a separate repo for ONNX weights is intended to be a temporary solution until WebML gains more traction. Converting models to ONNX using 🤗 Optimum is recommended for web-readiness, with ONNX weights located in a subfolder named 'onnx'.

LLM Name	Llama 68M
Repository 🤗	https://huggingface.co/Xenova/llama-68m
Base Model(s)	Llama 68M JackFram/llama-68m
Model Size	68m
Updated	2026-07-14
Maintainer	Xenova
Model Type	llama
Model Architecture	LlamaForCausalLM
Context Length	2048
Model Max Length	2048
Transformers Version	4.33.0.dev0
Tokenizer Class	LlamaTokenizer
Beginning of Sentence Token	<s>
End of Sentence Token	</s>
Unk Token	<unk>
Vocabulary Size	32000

Rank the Llama 68M 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 Llama 68M

Best Alternatives	Context / RAM	Downloads	Likes
Llama 68M Fft	4K / 0.1 GB	477	0
Llama 43M Beta	2K / 0.2 GB	6	0
Llama 68M Chat V1	2K / 0.3 GB	692	13
Llama 68M	2K / 0 GB	131563	38
Vicuna 68M	2K / 0.3 GB	11617	6
Gpt2 Sft Dutch	2K / 0.1 GB	9	0
Jack 68M Text Structurization	2K / 0.1 GB	10	2
Llama 68M Oasst Top1	2K / 0.3 GB	10	0
Llama 68M Oasst	2K / 0.3 GB	5	0
Llama 68M Finetuned Wikitext2	2K / 0.3 GB	10	0

Note: green Score (e.g. "73.2") means that the model is better than Xenova/llama-68m.