What are the hardware requirements for BOLETIN?

BOLETIN requires approximately 24.2 GB of VRAM. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed BOLETIN and how large is it?

BOLETIN is developed by bertin-project, a model with 6b 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 BOLETIN?

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

BOLETIN by bertin-project — VRAM 24.2GB

Name: BOLETIN
Author: bertin-project

BOLETIN is an open-source language model by bertin-project. Features: 6b LLM, VRAM: 24.2GB, License: apache-2.0, LLM Explorer Score: 0.09.

Arxiv:2509.24908 Base model:bertin-project/bert... Base model:finetune:bertin-pro... Endpoints compatible Es Gptj Pytorch Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/bertin-project/BOLETIN

BOLETIN Parameters and Internals

LLM Name	BOLETIN
Repository 🤗	https://huggingface.co/bertin-project/BOLETIN
Base Model(s)	Bertin GPT J 6B bertin-project/bertin-gpt-j-6B
Model Size	6b
Required VRAM	24.2 GB
Updated	2026-08-02
Maintainer	bertin-project
Model Type	gptj
Model Files	5.0 GB: 1-of-5 5.0 GB: 2-of-5 4.8 GB: 3-of-5 4.8 GB: 4-of-5 4.6 GB: 5-of-5 5.0 GB: 1-of-5 5.0 GB: 2-of-5 4.8 GB: 3-of-5 4.8 GB: 4-of-5 4.6 GB: 5-of-5
Supported Languages	es
Model Architecture	GPTJForCausalLM
License	apache-2.0
Model Max Length	2048
Transformers Version	4.34.0
Tokenizer Class	GPT2Tokenizer
Vocabulary Size	50400
Torch Data Type	float32
Activation Function	gelu_new
Errors	replace

Rank the BOLETIN 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 BOLETIN

Best Alternatives	Context / RAM	Downloads	Likes
Mlperf GPT J 6B	0K / 24.1 GB	11595	0
GPT J 6B	0K / 24.2 GB	246404	1526
Deception Normal	0K / 12.2 GB	8	0
Deception Filteredpositive	0K / 12.2 GB	7	0
Gptj Allenai Toxicity Blackbox	0K / 12.2 GB	9	0
Pygmalion 6B Roleplay	0K / 12.1 GB	782	2
...j Allenai Toxicity Explainable	0K / 12.2 GB	10	0
Pygmalion 6B	0K / 16.3 GB	286	753
Test GPT J 6B	0K / 2.5 GB	7	0
Explainable GPT J 6B	0K / 24.2 GB	5	0

Note: green Score (e.g. "73.2") means that the model is better than bertin-project/BOLETIN.