AI BriefingBig Enough
Key point
Mistral AI has unveiled Mistral Large 2, a high-performance model with 123 billion parameters.
Details
Mistral AI has unveiled Mistral Large 2, its next-generation model that pushes the boundaries of cost efficiency, speed, and performance.
This model supports a 128k context window, dozens of languages including Korean, and over 80 programming languages. It is designed with 123 billion (123B) parameters, enabling high-throughput inference on a single node.
Key features are as follows:
- Performance and Efficiency: It achieves 84.0% accuracy on MMLU, offering excellent performance-to-cost efficiency.
- Code and Reasoning: Trained on large-scale code data, it delivers performance on par with GPT-4o, Claude 3 Opus, and Llama 3 405B.
- Improved Reliability: It is trained to minimize hallucination and to recognize when information is insufficient, resulting in enhanced mathematical reasoning capabilities.
As for licensing, the Mistral Research License applies to research and non-commercial use, but a separate commercial license must be obtained for commercial self-deployment purposes.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.