Build with open-weight models

We release open-weight models for everyone to customize and deploy where they want it. Our super-efficient model Mistral Nemo is available under Apache 2.0, while Mistral Large 2 is available through both a free non-commercial license, and a commercial license.

• Download them for deployment in your own environment

• Use them on La Plateforme at market-leading availability, speed, and quality control

Cutting edge technology for developers.

Openness

We lead the market of open source generative technologies to bring trust and transparency in the field and foster decentralised technology development

Portability

Our technology is available through serverless APIs, public cloud services (on Azure AI, on Amazon Bedrock), and for VPC/on-premise deployment. Our independence from cloud providers is a guarantee of our customers’ independence.

Value and speed

Our flagship model, Mistral Large, has independently validated top-tier reasoning in multiple languages. All our models bring unmatched value and latency at their price points.

Customisation

Our models can be fine-tuned and modified at will for your business to create differentiated AI applications.