Microsoft signs agreement with Mistral AI from France, a competitor of OpenAI with its own chatbot
On Monday, Microsoft revealed a new partnership with the French startup Mistral AI in the field of artificial intelligence. This collaboration aims to reduce Microsoft’s dependence on OpenAI, the creator of ChatGPT, for the development of future chatbots and other generative AI technologies.
Mistral AI was born less than a year ago, but it was already described by Microsoft on Monday as an “innovator and pioneer” as a pioneer in building more efficient and cost-effective artificial intelligence systems.
Microsoft and Mistral did not disclose financial terms of the deal, although Microsoft said it would involve a small investment in the Paris-based startup. That suggests it’s much smaller than Microsoft’s billion-dollar investment in OpenAI, a years-long relationship that has drawn antitrust regulators in the United States and Europe.
Mistral also released a public test version of its own chatbot called Le Chat on Monday, which apparently was so flooded with interest that a company executive said it was temporarily down for part of Monday.
The company also announced its latest large language model, Mistral Large, which it claims is on par with competitors such as OpenAI’s GPT-4, Anthropic’s Claude 2 and Google’s Gemini Pro, and is available on Microsoft’s Azure cloud computing platform . Mistral has also previously said it will work with other major cloud service providers such as Amazon and Google.
Mistral made a splash by attracting large amounts of investor funding to give it a multibillion-dollar valuation just months after it was founded last spring. Its “open source” approach to AI development means it publicly releases the key components of its models, unlike companies like OpenAI, which closely guard them.
It was started by three French former Google and Meta researchers: CEO Arthur Mensch, Chief Scientist Guillaume Lample and CTO Timothee Lacroix.
As the European Union prepared the final version of its AI law, a comprehensive AI code, last fall, Mistral rejected efforts to impose restrictions on basic models using generative AI systems. Mensch said on social media that the EU’s proposals for a two-tier system would discourage innovative new entrants.
