Mistral AI, a Paris-based artificial intelligence startup, raised €385 million, or $415 million, in its second funding round in seven months, making it the latest company in the sector to achieve unicorn status.
The startup’s valuation soars to approximately $2 billion thanks to the funding, which was spearheaded by investors like Andreessen-Horowitz and LightSpeed Ventures. This puts it among the top AI companies globally. At a valuation of $260 million, Mistral AI raised about €105 million (roughly $113 million) in June.
The success of Mistral AI demonstrates the ongoing interest AI companies have sparked over the past 12 months. The reason for the industry’s success is the increased interest in generative AI tools like ChatGPT from OpenAI, Bard from Google, and Claude from Anthropic.
This trend was started by OpenAI, and many other AI companies expanded significantly by developing AI software or by acquiring AI startups.
With its newfound prominence, Mistral is ready to take on more established players like OpenAI and become a formidable European competitor. Arthur Mensch, the CEO and co-founder of Mistral AI, reportedly stated:
“Since the creation of Mistral AI in May, we have been pursuing a clear trajectory: that of creating a European champion with a global vocation in generative artificial intelligence, based on an open, responsible and decentralised approach to technology.”
Introduces a new product
Mistral AI opened beta access to its first platform service, “la plateforme,” during the height of its valuation boom. Three chat endpoints are available from this service, and they are made to produce text in response to text commands.
As per Mistral, “mistral-tiny” and “mistral-small” utilize publicly available open models, whereas “mistral-medium” functions using a prototype model that exhibits exceptional performance in real-world scenarios and is presently undergoing testing.
Furthermore, Mixtral 8x7B, an AI model that, according to Guillaume Lample, co-founder of the company, “matches or outperforms [Meta’s] Llama 2 70B and [OpenAI’s] GPT3.5 on most standard benchmarks,” was unveiled. It further stated that the “mistral-small” endpoint of Mixtral 8x7B is currently being used.
Mixtral 8x7B is accessible in beta access and was made available under the Apache 2.0 license.