HONG KONG (AP) — DeepSeek, a Chinese artificial intelligence firm, has stirred markets by claiming that its newest AI model, R1, rivals the capabilities of OpenAI's offerings, despite relying on less sophisticated computer chips and being more energy-efficient. This development has alarmed analysts as it suggests that China could be ahead of the U. S. in the AI competition, even with existing restrictions on its access to cutting-edge chips. DeepSeek is merely one among numerous Chinese enterprises dedicated to AI, all aiming to position China as the leading force in this sector by 2030, striving to surpass the U. S. in technological dominance. China, like the U. S. , is pouring billions into AI advancements. Notably, it announced a substantial 60 billion yuan ($8. 2 billion) AI investment fund just days after the U. S. enacted new chip export controls. Furthermore, Beijing is heavily investing in the semiconductor sector to enhance its own capabilities in producing sophisticated computer chips while navigating restrictions imposed by industry frontrunners. Companies are implementing talent programs and financial incentives and are set to launch AI academies and integrate AI education within the primary and secondary school systems. China has put in place regulations for AI, focusing on safety, privacy, and ethical considerations. The ruling Communist Party also dictates the topics that AI models can address: DeepSeek tailors its outputs to comply with these guidelines. Here is a brief look at some prominent AI models from China: **Alibaba Cloud’s Qwen-2. 5-1M** Alibaba Cloud’s Qwen-2. 5-1M represents the e-commerce giant’s open-source AI series. This model features large language capabilities, allowing it to handle lengthy inquiries and facilitate extensive and profound discussions.

Its proficiency in complex tasks like reasoning, dialogue, and code comprehension is steadily improving. In addition to its competitors, Alibaba Cloud has made Qwen, also known as Tongyi Qianwen in China, available to the public as a chatbot. The company’s array of AI solutions, including the Qwen2. 5 series, primarily serves developers and business clients in sectors such as automotive, banking, gaming, and retail, assisting in product development and enhancing customer interactions. **Baidu’s Ernie Bot 4. 0** Baidu, China's leading search engine, developed Ernie Bot, which was the first AI chatbot to be publicly accessible in China. Baidu claims that the public release of the model was aimed at gathering substantial real-world user feedback to enhance its functionalities. As of June 2024, Ernie Bot 4. 0 boasts over 300 million users. Like OpenAI’s ChatGPT, Ernie Bot enables users to pose questions and generate images in response to text prompts. **ByteDance’s Doubao 1. 5 Pro** ByteDance, the parent company of TikTok, launched the Doubao 1. 5 Pro AI model last week. Doubao has quickly gained popularity, amassing 60 million active users monthly. According to ByteDance, the Doubao 1. 5 Pro surpasses ChatGPT-4 in knowledge retention, coding, reasoning, and Chinese language processing. The company also asserts that this model is more cost-efficient, requiring less hardware investment than other large language models due to its optimized architecture that balances performance with reduced computational requirements. **Moonshot AI’s Kimi k1. 5** Beijing-based startup Moonshot AI, now valued at over $3 billion following its latest funding round, claims that its recently launched Kimi k1. 5 matches or even outperforms OpenAI’s o1 model. Kimi is designed to take additional time for thoughtful responses and tackle more challenging and intricate problems. Moonshot asserts that Kimi excels beyond OpenAI o1 in areas like mathematics, coding, and the ability to process both textual and visual inputs, including images and videos.