French AI company, Mistral AI, has entered the large language model (LLM) game with a bang, introducing Mistral Large. This new model boasts capabilities rivaling the top systems on the market, including OpenAI's recently launched GPT-4.

Mistral AI claims that Mistral Large outperformed most existing LLMs in language understanding, math, and coding tasks. Notably, it came close to matching the performance of GPT-4, considered the current leader. Additionally, Mistral Large offers multilingual capabilities, being fluent in English, French, Spanish, German, and Italian, and holding a vocabulary exceeding 20,000 words.

However, unlike Mistral's first open-source model, Mistral Large adopts a closed-source approach similar to its competitors like GPT-3 and GPT-4.

This move comes amidst a period of significant growth for Mistral AI. The company secured nearly $500 million in funding late last year, followed by a recent partnership with Microsoft. This partnership allows access to Mistral Large through Azure cloud services, potentially expanding its reach and impact.

However, a potential hurdle lies ahead. Microsoft's €15 million investment in Mistral AI has drawn scrutiny from EU regulators, concerned about the tech giant's existing ties to OpenAI. This review could potentially affect the Mistral-Microsoft partnership.

Despite the potential challenges, Mistral Large marks a significant step forward for the French startup. With competitive pricing starting at $8 per million input tokens and $24 per million output tokens, Mistral AI positions itself as a serious contender in the increasingly crowded LLM landscape.

While its full potential remains to be seen, Mistral Large is undoubtedly a strong entry and could shake up the established hierarchy in the AI space.