Deepseek, a China-based artificial intelligence company, has announced an upgrade to an AI chatbot. It says it has reduced hallucination rates and enhanced overall logic, mathematics and programming.
According to DeepSeek, the upgraded model, DeepSeek-R1-0528, “has greatly improved the depth of inference and inference capabilities.” The startup said the overall performance of the model is “closing the performance of major models, such as the O3 and Gemini 2.5 Pro.”
Deepseek’s R1 chatbot debut in January sent shockwaves through the AI industry, establishing more China as an AI force. The company’s first AI model had a training cost of $6 million, which performed similarly to the major AI models trained with significantly larger capital.
According to Apps Business data, Deepseek has been downloaded 75 million times since its launch and had 38 million active users (MAUs) per month as of April. In a recent anti-trust lawsuit, Google estimated that Gemini reached 350 million active users in March, while Openai’s ChatGPT charged 600 million active users in the same month.
Related: China’s Deepseek launches new open source AI after R1 joins Openai
https://www.youtube.com/watch?v=fykm3ka_lzu
Chinese American AI races get hot
The US government plans to limit the sale of advanced chip design software to China. The move seeks to limit China’s ability to advance its domestic semiconductor manufacturing capacity, according to a Bloomberg report.
Semiconductors are important for a wide range of technologies, including AI, and act as a hardware backbone for training and running complex models.
New Chinese AI models, such as Tencent’s T1 and Alibaba’s QWEN3, have also spurred AI racing in the first few months of 2025.
Magazine: AI Eye: 9 Curious Things About Deepseek R1
