Mistral Models

Mistral AI's model family including Mistral Large, Mistral Small, Mixtral 8x7B, and Codestral delivers high performance relative to parameter count through efficient architecture choices including mixture-of-experts design that enables frontier-quality results at significantly lower inference cost than comparable dense models. Its open-weight Mixtral models are among the most capable freely available models for commercial use, making Mistral a leading choice for cost-sensitive production deployments. Developers and enterprises building high-throughput AI applications use Mistral models for their combination of strong benchmark performance, efficient inference economics, and the option of open-weight deployment.