Liang Wenfeng, founder of the Chinese AI startup DeepSeek, attended a government symposium hosted by Premier Li Qiang in Beijing on Jan. 19.

Wenfeng has been recognized as the “biggest dark horse” of 2025 in the open-source large language model (LLM) sector, establishing himself as a new prominent figure in China’s AI industry.

The event gathered experts from various sectors to discuss a draft government work report, according to the official news agency Xinhua.

DeepSeek, located in Hangzhou, gained attention in December 2024 when its DeepSeek V3 large language model (LLM) performed well in benchmark tests.

The model, despite limited resources, outperformed those from major companies like Meta Platforms and OpenAI, as reported by industry sources.