Mistral AI Unveils Mistral Large: A New Contender in the Large Language Model Arena
Mistral AI, a France-based startup, has recently made waves in the AI community with the introduction of their latest large language model (LLM) called Mistral Large. This new model is positioned to rival major players in the industry, boasting impressive capabilities and performance.
According to Mistral AI, Mistral Large has demonstrated exceptional language understanding abilities, outscoring many top LLMs in various tests. Particularly noteworthy is its performance compared to OpenAI’s GPT-4, where Mistral Large showcased remarkable proficiency not only in language comprehension but also in mathematical and coding assessments.
Co-founder and Chief Scientist of Mistral AI, Guillaume Lample, emphasized that Mistral Large signifies a significant leap forward from the company’s previous models. In addition to the model itself, Mistral AI has also launched a user-friendly chatbot interface named Le Chat, enabling seamless interaction with the system akin to ChatGPT.
One of the key strengths of Mistral Large lies in its multilingual capabilities, with fluency in English, French, Spanish, German, and Italian, supported by a vast vocabulary exceeding 20,000 words. While Mistral AI’s earlier models were open-source, Mistral Large’s code is now closed, aligning with the approach taken by industry leaders like OpenAI.
The substantial funding injection of nearly $500 million received by Mistral AI from prominent backers such as Nvidia and Andreessen Horowitz underscores the industry’s confidence in the startup’s potential. Moreover, the recent collaboration with tech giant Microsoft to integrate Mistral Large into Azure cloud services further solidifies Mistral AI’s position in the AI landscape.
However, Microsoft’s significant investment in Mistral AI has raised eyebrows among European Union regulators, who are closely monitoring the tech giant’s ties to key players like OpenAI. The European Commission’s scrutiny of this partnership could potentially impact the future trajectory of Mistral AI and its collaborations.
Pricing details reveal that access to Mistral Large starts at $8 per million input tokens and $24 per million output tokens, leveraging Azure’s robust computing infrastructure for training and deployment requirements. Despite awaiting comprehensive third-party evaluations, Mistral AI’s track record with Mistral Medium, ranking 6th among over 60 language models, positions the company as a formidable contender in the competitive AI landscape.
In conclusion, Mistral AI’s unveiling of Mistral Large marks a significant milestone in the evolution of large language models, setting the stage for increased competition and innovation in the AI domain.
Image Source: Unsplash
