Global AI and Data Science

 View Only

🚀 Breaking News: Mistral Large 2 is here!

By MARYAM ASHOORI posted 2 days ago

  
This has been such an amazing week with our partners dropping powerful flagship models. 🌟

I’m thrilled to announce that Mistral Large 2, the most powerful Mistral model, is now available at watsonx.ai for enterprise use! 🎉

Mistral Large 2 is a cutting-edge LLM that excels in reasoning, knowledge, and coding domains.

Here’s what you need to know:

🌐 Available on watsonx.ai
🧠 123B parameters, large enough to fit on a single-node (larger throughput for its performance scale)
🌍 Strong Multilingual support to handle dozens of languages, including French, German, Spanish, Italian, Portuguese, Arabic, Hindi, Russian, Chinese, Japanese, and Korean
💻 Trained on 80+ coding languages including Python, Java, C, C++, JavaScript, and Bash. Performance accuracy of 74% on code generation, higher than Llama 3.1 and close to GPT-4o
🏆 Achieves 84% on MMLU; 8.63 MT Bench; 92% on HumanEval; on par with leading models such as GPT-4o, Claude 3 Opus, and Llama 3 405B
🔓 Open model, released under the Mistral Research License
🔍 Fine-tuned to minimize “hallucination” (more cautious and discerning in its responses to ensure reliable and accurate outputs)
📝 Tuned to follow precise instructions and handling long multi-turn conversations
🔧 Native function calling and JSON outputting features, delivering higher accuracy compared to GPT4o, Claude 3 Opus, and 3.5 Sonnet
🧮 Advanced Math and Reasoning capabilities


Check out the official release announcement:
https://lnkd.in/eJtRBSvK

Congratulations to our friends and partners at Mistral Eléonore Arcelin Gianna Maria Lengyel Margaret J.

(Note: this post was written with the assistance of Mistral Large 2 on watsonx.ai!)
Image preview
0 comments
0 views

Permalink