Maistros: A Greek Large Language Model Adapted Through Knowledge Distillation From Large Reasoning Models

Nikolaos Giarelis, Charalampos Mastrokostas, Nikos Karacapilidis · 2026

Read Maistros: A Greek Large Language Model Adapted Through Knowledge Distillation From Large Reasoning Models by Nikolaos Giarelis, Charalampos Mastrokostas, Nikos Karacapilidis (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-01870.txt.

Description

Nikolaos Giarelis, Charalampos Mastrokostas, and Nikos Karacapilidis focus on Modern Greek, a language with little training data for large language models. Big reasoning models can answer hard questions well, but they are huge and slow. The team asks whether their skill can be passed to a smaller Greek model that runs on normal hardware.

They build CulturaQA, a Greek question and answer set made with a large reasoning model and checked by humans. Using it, they train Maistros, an open 8-billion-parameter model, by distillation and fine-tuning. They also give a light method for testing models and compare nine models on nine Greek test sets.

Catalog record

CREATORNikolaos Giarelis, Charalampos Mastrokostas, Nikos Karacapilidis
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE49 KB
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.01870

Browse the full library · About the Tower of Babel