Magistral is Mistral AI's first reasoning model, excelling in domain-specific, transparent, and multilingual reasoning. Available as an open-source 24B version (Small) and a powerful enterprise version (Medium).
Mistral's models have now officially joined the reasoning lineup with their new release, Magistral. It's a model family focused on transparent, multi-step logic for tackling complex problems.
Magistral comes in two versions: a powerful enterprise "Medium" model and a 24B "Small" model. The Small version is an open-weight model under the Apache 2.0 license, which is super awesome for the community to build upon.
It's designed to show its work, providing a traceable thought process in multiple languages. Performance also looks strong, with impressive scores on math benchmarks and up to 10x faster responses in Le Chat.
