Skip to content

Pleias

Building frontier AI for document and data processing in regulated industries.

Overview

Github
Code

More Information

Pleias models are:

  • Compliant with EU legislation. They have open weights, open code, and open data without any copyright or GDPR-related concerns.
  • Energy efficient. Deployable on consumer GPUs and CPUs.
  • Excel at multilingual use. Unlike other small language models, Pleias models excel at many European languages. Vision models are trained on European documents, which different considerably from typical American data used by other providers.

Contribute

Join the Discord

Contributors

About Anastasia

Anastasia is a Co-Founder of Pleias, a French startup specializing in developing open science LLMs trained on fully open copyright-free data. She is also s Senior Associate Lecturer in Data Analysis and Digital Strategy at Sorbonne-Nouvelle University.

More from Anastasia

About Pierre-Carl

Pierre-Carl is the Co-Founder of Pleias. He was previously Head of Research at opsci. He received a PhD in Information and Communication from the Sorbonne, for which he was awarded the

More from Pierre-Carl

Other Projects