Mistral AI

Based in France, Mistral AI is an open-source GenAI company that develops fine-tuned language models and provides businesses with foundational technologies to automate existing processes and develop new products.

The company's flagship model, Mistral-7B, supports English text and code generation abilities and is positioned as a cost-effective alternative to larger LLMs, providing similar capabilities at a reduced compute cost. The company aspires to support the open GenAI community and emphasizes its commitment to achieving state-of-the-art performance with open models. Mistral's business model involves offering a free downloadable model while providing commercial offerings as white-box solutions, with both weights and code sources accessible, targeting enterprises with hosted solutions and dedicated deployments.

Mistral-7B is available for customers through Amazon SageMaker JumpStart by Amazon Web Services (AWS) and Google's Vertex AI platform.

In February 2024, Mistral AI introduced an LLM, Mistral Large, for reasoning capabilities. It supports English, French, Spanish, German, and Italian languages. Moreover, in July 2024, the company released Mistral Large 2, supporting a diverse range of languages and coding languages, with a 128,000-token context window and 123 billion parameters.

In April 2024, the company released Mistral 8x22B, an LLM built on a mixture of expert (MoE) architecture with 176 billion parameters. The model is available under a permissive Apache 2.0 license.

In May 2024, Mistral AI, released Codestral, a model for coding with 22 billion parameters and programmed across 80 programming languages. It also released Codestral Mamba for faster and longer code generation, Mathstral for math-related reasoning and scientific discovery, a small language model named Mistral NeMo and in July 2024.

In September 2024, the company released Pixtral 12 billion, its first multimodal AI model with language and vision processing capabilities, followed by Pixtral Large with enhanced text and visual data processing in November 2024.

In November 2024, Mistral AI launched a content moderation API that powers its Le Chat service. The API is designed to help users implement content moderation in their applications according to their safety standards.

Key customers and partnerships

In December 2023, Mistral AI partnered with Google Cloud to distribute its LLMs globally using Google's AI-optimized infrastructure for developers and businesses.

In January 2024, the company partnered with Brave, a provider of a browser-based AI chatbot named Leo, to introduce the updated Mistral AI's Mixtral 8x7B as the default LLM.

Mistral AI formed several partnerships in February 2024 to introduce its foundation models to its partners' platform. These included Capgemini, a provider of technology and consulting, Amazon, Microsoft, and IBM to integrate Mistral 7B and Mixtral 8x7B into Amazon Bedrock, Microsoft's Azure, and watsonx cloud platforms, repectively. In March 2024, the company partnered with Snowflake, a data management and analytics platform, to provide Mistral Large in the Snowflake data cloud.

In April 2024, Mistral partnered with Amazon to introduce Mistral Large to Amazon Bedrock.

In June 2024, the company partnered with SAP to offer access to its latest models.

In July 2024, the company partnered with Dassault Systemes to expand the latter’s product portfolio and with BNP Paribas, a global bank to provide its large language models (LLMs).