Tencent Officially Launches T1 Reasoning Model in China's AI Race

The T1 model, powered by the Turbo S foundational language model, claims faster query processing and outperforms DeepSeek on key benchmarks.

Tencent's T1 reasoning model, launched on March 21, 2025, is designed for faster responses and processing extended text documents with minimal hallucination rates.
The T1 model is powered by Tencent's Turbo S foundational language model, unveiled in February 2025, and reportedly processes queries faster than DeepSeek's R1 model.
Tencent released a chart showing the T1 model outperforming DeepSeek's R1 on some knowledge and reasoning benchmarks.
Tencent plans to increase AI investments in 2025 after significant spending throughout 2024, signaling its commitment to advancing AI technologies.
Beijing continues to support domestic AI innovation, including Manus, which recently cleared regulatory hurdles for its AI assistant Monica and gained state media attention.