- Mistral has unveiled its new flagship AI model, Large 2, which it claims rivals the latest models from OpenAI and Meta in terms of code generation, maths and reasoning.
- Despite high costs and legal challenges, Large 2 demonstrates significant improvements in multilingual support and response accuracy.
OUR TAKE
Mistral’s Large 2, though with fewer parameters (123 billion), it offers superior performance in these areas compared to Meta’s recent launch. Mistral focuses on reducing Large 2’s hallucination problems, training it to be more insightful in its responses. The company, which recently raised $640 million at a $6 billion valuation, is quickly becoming a significant player in AI. However, like most AI models, Large 2 is not free for commercial use and requires a paid license.
-Rae Li, BTW reporter
What happened
Paris-based AI startup Mistral on Wednesday unveiled a new flagship AI model called Large 2, which is claimed to match the latest cutting-edge models from OpenAI and Meta in terms of code generation, maths and reasoning. Mistral claims that Large 2 sets a new standard for open source models in terms of performance and cost-effectiveness. In terms of code generation and mathematical performance, Large 2 appears to outperform Llama 3.1.
During training, Mistral pays particular attention to reducing the model’s illusion problem by training Large 2 to be able to admit ignorance when it does not know the answer, rather than making up what appears to be reasonable. It is vital to note that Mistral’s models are not open source in the traditional sense, and any commercial application requires a paid licence. Also, due to the sheer scale of the model, only a few people in the world have the expertise and infrastructure to implement such large models. Mistral Large 2 is available on Google Vertex AI, Amazon Bedrock, Azure AI Studio, and IBM watsonx.ai.
Also read: Meta launches an enhanced version of its Llama 3 AI model
Also read: Meta releases powerful Llama 3.1 AI model for free, sparking debate
Why it’s important
The release of Mistral’s new AI model, Large 2, marks a new breakthrough in the field of AI. This development not only highlights the rapid development of AI technology, but also reflects the fierce competition within the industry.
The introduction of the Large 2 and its efforts to reduce model illusion problems demonstrates the advances in accuracy and reliability of AI models. Nevertheless, its non-open source business model and high requirements for expertise and infrastructure also reveal the challenges of AI technology adoption and application.
