What are the hardware requirements for Quokka 1.3B?

Quokka 1.3B requires approximately 2.7 GB of VRAM. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Quokka 1.3B and how large is it?

Quokka 1.3B is developed by Corianas, a model with 1.3b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

How does Quokka 1.3B perform on standard benchmarks?

Quokka 1.3B has the following published scores: MMLU 26.66. Compare against reference models on this page or on the LLM Explorer leaderboards.

Quokka 1.3B by Corianas — VRAM 2.7GB

Name: Quokka 1.3B
Author: Corianas

Quokka 1.3B is an open-source language model by Corianas. Features: 1.3b LLM, VRAM: 2.7GB, License: apache-2.0, LLM Explorer Score: 0.11, Arc: 27.7, HellaSwag: 37.9, MMLU: 26.7.

Arxiv:1910.09700 Dataset:guanaco/guanaco Dataset:the pile En Endpoints compatible Gpt2 Pytorch Region:us Safetensors Tensorboard

Model Card on HF 🤗: https://huggingface.co/Corianas/Quokka_1.3b

Quokka 1.3B Parameters and Internals

Use Cases

Primary Use Cases:

GPT chatbot that can follow instructions

Limitations:

Not suitable for accurate use, do not use for medical or financial advice

Additional Notes

Has limitations in knowledge about facts, better suited for understanding instructions than delivering accurate information.

Supported Languages

en (NLP)

Training Details

Data Sources:

the_pile, guanaco/guanaco

Methodology:

Finetuned using DataBricks Dolly Framework

Hardware Used:

8xA100s

LLM Name	Quokka 1.3B
Repository 🤗	https://huggingface.co/Corianas/Quokka_1.3b
Model Size	1.3b
Required VRAM	2.7 GB
Updated	2026-09-09
Maintainer	Corianas
Model Type	gpt2
Model Files	2.7 GB 2.7 GB 0.0 GB
Supported Languages	en
Model Architecture	GPT2LMHeadModel
License	apache-2.0
Transformers Version	4.25.1
Tokenizer Class	GPT2Tokenizer
Beginning of Sentence Token	<\|endoftext\|>
End of Sentence Token	<\|endoftext\|>
Unk Token	<\|endoftext\|>
Vocabulary Size	50260
Torch Data Type	bfloat16
Activation Function	gelu
Errors	replace

Rank the Quokka 1.3B 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 Quokka 1.3B

Best Alternatives	Context / RAM	Downloads	Likes
Novi Nano Base	0K / 0 GB	167	1
Cerebras 1.3b Quantized	0K / 1.4 GB	66	0
GPT Sw3 1.3B Instruct	0K / 5.5 GB	1068	3
1.3B	0K / 2.7 GB	80	2
GPT Sw3 1.3B	0K / 5.5 GB	42	6
MGPT 1.3B Kazakh	0K / 5.8 GB	21	8
LaMini Cerebras 1.3B	0K / 5.4 GB	90	3

Note: green Score (e.g. "73.2") means that the model is better than Corianas/Quokka_1.3b.