#DeepSeekImpact

DeepSeek is a Chinese artificial intelligence (AI) company that has attracted attention with the launch of its latest open source AI model, DeepSeek v3, in December 2024. The model has 671 billion parameters, making it the largest open source AI model to date, surpassing Llama 3.1 which had 405 billion parameters.

DeepSeek v3 has shown superior performance in various tests. In the Math-500 mathematics test, the model scored 90.2, outperforming the Qwen model which scored 80. In addition, in a programming competition on the Codeforces platform, DeepSeek v3 managed to beat other models such as OpenAI's GPT-4o and Alibaba's Qwen 2.5.

However, along with its superiority, a security vulnerability was also discovered in the DeepSeek chatbot related to prompt injection attacks. This vulnerability allows attackers to take over user accounts by manipulating the input received by the AI. DeepSeek has released an update to address this issue, emphasizing the importance of awareness and proactive action in maintaining the security of AI systems.

With the launch of DeepSeek v3, the company offers a powerful open source alternative in the AI ​​space, providing developers and researchers with the opportunity to leverage cutting-edge technology with more open access.