- Paris-based Mistral AI has introduced its latest AI model, Large 2, boasting 123 billion parameters and superior capabilities in code generation, mathematics, and reasoning.
- The model aims to address hallucination issues and supports multiple languages, positioning it ahead of competitors like GPT-4 and Llama 3.
Large 2 Unveiling: It’s a New Benchmark in AI.
The latest AI model by the Paris-based startup Mistral AI is Large 2. This new model has 123 billion parameters, which will likely make it outperform other notable competitors like OpenAI’s GPT-4, Anthropic’s Claude 3, and Meta’s Llama 3.1, which have very few parameters compared to it. Mistral claims that this marks a significant improvement over their previous model, Mistral Large, as they state that it excels in certain domains such as mathematics, code generation, and reasoning.
Read more : Unlocking Industrial Potential: Cisco and Rockwell’s APAC Partnership
Focus on Reducing Hallucination Issues
Mistral AI has focused heavily on reducing hallucination problems while developing Large 2 because hallucination is pervasive across AI models where the system makes up responses. The company ensures Large 2 is more cautious and accurate with its outputs. More so, the attention to details improves its reliability, making it more trustworthy to users as well.
Enhanced Multilingual Support and Large Token Window
With a token window of enormous magnitude i.e.,128000 tokens, large two can process long prompts like those of an entire text-book running into several hundred pages. This capability puts it ahead of many other existing models. Further still, there are other multilingual supports offered by large two such as English, French, German, Spanish, Italian, Portuguese, Arabic, Hindi, Russian, Chinese, Japanese, and Korean, among many more, in addition to support for eighty coding languages… However, Large Two isn’t multimodal, unlike Google’s Gemini models and OpenAI’s GPT-4.
Conclusion
The incredible 123 billion parameters of the Mistral AI’s Large 2 model together with improved capabilities in math coding multilingual support sets a new standard in the industry about Artificial Intelligence (AI). While addressing crucial issues such as hallucination; the truth that this does not stop being multimodal shows how much room for growth remains available again. Backed by significant funding and a robust feature set, Large 2 is positioned to revolutionize the AI landscape, outpacing competition and pushing the limits of AI capabilities.
Latest Stories :
CrowdStrike Update Sparks Unprecedented Global IT Chaos
Meta clarifies AI content: “Made by AI” evolves to “AI Info”
OpenAI’s Bold Step: Developing In-House AI Chips
What is your reaction?
