What are the hardware requirements for Garrulus?

Garrulus requires approximately 14.4 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 Garrulus and how large is it?

Garrulus is developed by udkai, a model with 7b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

How does Garrulus perform on standard benchmarks?

Garrulus has the following published scores: MMLU 64.57. Compare against reference models on this page or on the LLM Explorer leaderboards.

Garrulus by udkai — VRAM 14.4GB, 32K context

Name: Garrulus
Author: udkai

Garrulus is an open-source language model by udkai. Features: 7b LLM, VRAM: 14.4GB, Context: 32K, License: apache-2.0, LLM Explorer Score: 0.11, Arc: 73.3, HellaSwag: 88.9, MMLU: 64.6, GSM8K: 64.5.

7b Base model:finetune:mlabonne/n... Base model:mlabonne/neuralmarc... Dataset:hromi/winograd dpo bas... Deploy:azure Doi:10.57967/hf/1590 Dpo Endpoints compatible Mistral Mlabonne/neuralmarcoro14-7b Region:us Safetensors Sharded Tensorflow Winograd

Model Card on HF 🤗: https://huggingface.co/udkai/Garrulus

Garrulus Parameters and Internals

Model Type

CAUSAL_LM

Additional Notes

The model has been intentionally contaminated with two epochs of DPO, leading to improved performance on the Winogrande dataset as well as other metrics like TruthfulQA, HellaSwag, and ARC challenge.

Training Details

Data Sources:

hromi/winograd_dpo_basic

Methodology:

Direct Preference Optimization (DPO) with the Winograd dataset.

Hardware Used:

A40 GPU

LLM Name	Garrulus
Repository 🤗	https://huggingface.co/udkai/Garrulus
Base Model(s)	NeuralMarcoro14 7B mlabonne/NeuralMarcoro14-7B
Model Size	7b
Required VRAM	14.4 GB
Updated	2026-07-17
Maintainer	udkai
Model Type	mistral
Model Files	4.9 GB: 1-of-3 5.0 GB: 2-of-3 4.5 GB: 3-of-3
Model Architecture	MistralForCausalLM
License	apache-2.0
Context Length	32768
Model Max Length	32768
Transformers Version	4.37.0.dev0
Tokenizer Class	LlamaTokenizer
Padding Token	<unk>
Vocabulary Size	32000
Torch Data Type	bfloat16

Rank the Garrulus 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

Quantized Models of the Garrulus

Model	Likes	Downloads	VRAM
Garrulus GGUF	8	400	2 GB
Garrulus GPTQ	3	28	4 GB
Garrulus AWQ	2	7	4 GB

Best Alternatives to Garrulus

Best Alternatives	Context / RAM	Downloads	Likes
...Nemo Instruct 2407 Abliterated	1000K / 24.5 GB	220	20
MegaBeam Mistral 7B 512K	512K / 14.4 GB	8470	54
SpydazWeb AI HumanAI RP	512K / 14.4 GB	16	1
SpydazWeb AI HumanAI 002	512K / 14.4 GB	18	1
...daz Web AI ChatML 512K Project	512K / 14.5 GB	12	0
MegaBeam Mistral 7B 300K	282K / 14.4 GB	3779	16
MegaBeam Mistral 7B 300K	282K / 14.4 GB	8018	17
Hebrew Mistral 7B 200K	256K / 30 GB	23	15
Astral 256K 7B	250K / 14.4 GB	9	0
Astral 256K 7B V2	250K / 14.4 GB	8	0

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