Mistral AI is a French artificial intelligence company specializing in developing open-source large language models (LLMs). Founded in April 2023 by former researchers from Meta Platforms and Google DeepMind, the company is headquartered in Paris.
Released in September 2023, this 7.3-billion parameter model outperforms Meta's LLaMA 2 13B on various benchmarks, including code and reasoning tasks. It utilizes grouped-query and sliding window attention mechanisms for efficient inference and can be fine-tuned for chat and other applications.
Introduced in December 2023, this model employs a sparse mixture of experts architecture, comprising eight distinct groups of "experts," resulting in a total of 46.7 billion usable parameters. Each token utilizes 12.9 billion parameters, balancing performance with computational efficiency.
Launched in July 2024, this model features 123 billion parameters and a context length of 128,000 tokens. It is fluent in multiple languages, including various programming languages, and is competitive with Meta's Llama 3.1 405B, particularly in programming-related tasks.
In June 2024, Mistral AI secured €600 million ($645 million) in a funding round led by General Catalyst, with participation from existing investors. This investment elevated the company's valuation to €5.8 billion ($6.2 billion), positioning it as a leading AI startup in Europe.
Mistral AI aims to democratize artificial intelligence by focusing on open-source innovation, providing accessible and customizable AI models for developers and businesses. Its rapid growth and significant funding underscore Europe's potential in the global AI landscape.
No comments yet. Be the first to comment!