Arthur Mensch is co-founder and CEO of Mistral AI, the Paris-based company that became Europe’s most prominent large-language-model developer. Before founding Mistral he was a researcher at DeepMind in Paris, where he worked on large language models and contributed to influential projects, including the Chinchilla scaling study that reshaped how labs trade off model size against training data.
In 2023 Mensch left DeepMind to start Mistral AI with Guillaume Lample and Timothee Lacroix, two former Meta researchers. The company, founded in April 2023, quickly drew attention for releasing capable models with open weights, beginning with Mistral 7B. Mistral’s own about page describes its mission as making frontier AI open to all and positions the firm as a European-led alternative built on openness, transparency, and returning control to users, in deliberate contrast to companies it saw closing off their work.
Mensch has become a leading voice for the argument that frontier AI should not be controlled by a small number of American companies, and that open-weight models support sovereignty and competition. For a general reader, he illustrates how quickly a new lab can reach the frontier and how the open-versus-closed debate over model weights became central to AI strategy and policy.