Mistral

Mistral develops large language models and the tools to put them to work inside companies. Developers reach its models through an API and through Studio, where they build, test and run AI agents and applications. Forge lets an enterprise train, align and evaluate custom models on its own data, and Compute provides dedicated GPU clusters for training and inference. Vibe is its agent for long, multi-step tasks, with a coding version that works in the terminal and the IDE. Several models are released with open weights and can be deployed on a company's own infrastructure. It is used by developers and by enterprises that want control over where their data goes.