China's DeepSeek AI outperforms ChatGPT

Innovations in Artificial Intelligence (AI) are emerging globally. Recently, China's new AI model, DeepSeek, has garnered significant attention as it surpasses the capabilities of well-known AI systems like ChatGPT, Gemini, and Claude AI.

The new AI DeepSeek remains popular on social media and the stock market. In layman's terms, DeepSeek involves conducting research, addressing complex questions, and making enhancements.

Founder of DeepSeek

Liang Wenfeng developed the renowned AI model DeepSeek in Hangzhou. According to Bernstein's report, DeepSeek has produced two primary AI models.

'DeepSeek V3' and 'DeepSeek R1' V3 models are advanced AI systems utilizing a Mixer-of-Experts (MOE) architecture. This method integrates multiple smaller models to work in unison, achieving superior performance with considerably fewer computing resources compared to other large models. The V3 model boasts a total of 671 billion parameters.

It also includes cutting-edge technologies like Multi-Head Latent Attention (MHLA) to lower memory consumption and mixed-precision training with FP8 computation to enhance efficiency.