France-based startup Mistral AI has unveiled its latest proprietary large language model (LLM), Mistral Large, entering a competitive AI market with claims of superiority over several major competitors.
Mistral Large reportedly outperformed several notable LLMs in a “multitask language understanding” test, demonstrating strong performance in math and coding evaluations. However, it fell short compared to GPT-4 in the same assessment.
While Mistral AI touted Mistral Large’s performance, it did not directly compare it to recent competitors such as xAI’s Grok and Google’s Gemini Ultra. Notably, these models were released in November and February, respectively.
Guillaume Lample, Mistral AI’s founder and chief scientist, highlighted Mistral Large’s significant improvement over the company’s earlier models. Additionally, Mistral AI introduced “Le Chat,” an AI chat interface akin to ChatGPT built on top of its models.
Mistral AI secured $487 million in funding in December from notable investors like Nvidia, Salesforce, and Andreessen Horowitz. The firm also announced a partnership with Microsoft, making Mistral Large accessible on Azure AI Studio and Azure Machine Learning.
Mistral Large, a closed, proprietary model, offers support for multiple languages and costs $8 per one million tokens of input and $24 per million tokens of output. Its valuation reportedly reached nearly $2 billion in December, according to Bloomberg.
While Mistral AI’s earlier model, Mistral Medium, ranks sixth among over 60 LLMs on platforms like Chatbot Arena, some observers on X expressed disappointment over Mistral Large’s shift to a closed, proprietary model.
Mistral AI and Microsoft plan to collaborate on AI research and development, leveraging Azure’s supercomputing infrastructure for training and scaling Mistral Large. This partnership marks a significant step forward in Mistral AI’s commercial-focused LLM endeavors.
Get $200 Free Bitcoins every hour! No Deposit No Credit Card required. Sign Up