DeepSeek has launched the new version of its V3 large language model, DeepSeek-V3-0324. This update significantly enhances the model's logical reasoning, programming, and mathematical problem-solving abilities. With this move, the company aims to strengthen its competition with leading firms like OpenAI and Anthropic.
The DeepSeek-V3-0324 model consists of 685 billion parameters, slightly more than its previous version. This increase allows the model to handle more complex tasks. The model also stands out with optimized front-end web development and improved Chinese language processing capabilities. Unlike previous versions, the new model is released under the MIT license, offering developers the freedom to use it for both commercial and non-commercial purposes.
The model's capabilities in programming and solving mathematical problems have been notably improved. Additionally, it demonstrates enhanced performance in processing Chinese texts and programming tasks.
DeepSeek’s more affordable and powerful model not only competes with OpenAI and other Western companies but also aims to reduce costs in the AI field and make the technology more accessible.
