Magistral: Efficient Training of Small Language Models for Reasoning

Mistral AI's about page lists Guillaume Lample as one of the company's three founders. His OpenReview profile lists expertise in machine translation and natural language processing and a PhD in computer science at Universite Pierre et Marie Curie - Paris 6.

AI Scientist & Observability Lead at Mistral AI.

Research scientist at Mistral AI working on language models for code.

Researcher and engineer whose public work focuses on large language models, reasoning systems, and open-source machine learning.

Marc'Aurelio Ranzato is listed as an author of the Mistral AI technical report Magistral: Efficient Training of Small Language Models for Reasoning.

Research scientist at Mistral AI working on deep learning, language technologies, and large-scale AI systems.

Julien Besombes is listed as an author of the Mistral AI technical report Magistral: Efficient Training of Small Language Models for Reasoning.

Aleksandra Piktus is a research scientist at Mistral AI. Her public speaker profiles say she works on multilinguality, representation learning, and responsible AI, and previously worked at the University of Cambridge on multimodal representation learning and cross-lingual transfer, where she completed a PhD.

Alexandre Rame is a research scientist at Mistral AI working on post-training, evaluation, and reasoning for large language models.

Saffet Kaya is listed as an author of the Mistral AI technical report Magistral: Efficient Training of Small Language Models for Reasoning.

Canonical link

Guillaume Lample

Pierre Stock

Theophile Gervet

Thomas Wolf

Marc'Aurelio Ranzato

Clement Gehring

Julien Besombes

Aleksandra Piktus

Alexandre Rame

Saffet Kaya