Chinese tech giant Tencent unveiled the official version of its T1 reasoning model on Friday night, escalating competition in China’s rapidly growing artificial intelligence sector.
The enhanced T1 model promises faster response times and improved performance on long-text processing tasks, Tencent announced via its official WeChat account. The company emphasized that T1 maintains “clear content logic and clean text” while keeping hallucination rates “extremely low”—a “key challenge in AI development.
The launch follows increased rivalry in China’s AI landscape, particularly after DeepSeek introduced models boasting performance on par with, or better than, Western alternatives at significantly lower costs.
Tencent had previously released a preview version of T1 through various platforms, including its AI assistant app Yuanbao. The newly launched official version runs on Tencent’s Turbo S foundational language model, unveiled last month. According to the company, Turbo S processes queries faster than DeepSeek’s competing R1 model.
A performance chart shared in Tencent’s announcement showed T1 outperforming DeepSeek R1 on select knowledge and reasoning benchmarks—a clear sign of Tencent’s push to dominate the space.
The company has ramped up AI investments throughout 2024 and plans to further increase capital expenditures in 2025, signaling its commitment to advancing AI technologies.
