Mistral releases Large 2, its answer to Meta and OpenAI’s latest models
Mistral, a Paris-based AI startup, has released its new flagship model, Large 2, which it claims is on par with the latest cutting-edge models from OpenAI and Meta in terms of code generation, mathematics, and reasoning. This release comes just one day after Meta dropped its latest open-source model, Llama 3.1 405b.
Mistral Large 2 boasts impressive performance, outpacing Llama 3.1 405b on code generation and math performance while using under a third of the parameters, with 123 billion parameters. The company focused on minimizing hallucination issues during training, ensuring the model is more discerning in its responses and acknowledges when it does not know something instead of making something up.
Large 2 is available for use on various platforms, including Google Vertex AI, Amazon Bedrock, Azure AI Studio, and IBM watsonx.ai. It also includes improved multilingual support, understanding 12 languages and 80 coding languages. Notably, the model produces more concise responses than leading AI models, which often provide lengthy responses.
Mistral recently raised $640 million in a Series B funding round, led by General Catalyst, at a $6 billion valuation. While Mistral’s models are not open source in the traditional sense, requiring a paid license for commercial use, the company is quickly shipping AI models that are near the cutting edge.
