Tencent has officially unveiled its T1 reasoning model, marking a significant step forward in the increasingly competitive artificial intelligence sector in China.

Announced on 21 March 2025, the latest T1 model delivers faster response times and advanced capabilities, particularly for processing extended text documents, according to Tencent’s official WeChat post.

The T1 reasoning model is designed to maintain clear content logic and polished text, boasting an exceptionally low hallucination rate. Powered by Tencent’s Turbo S foundational language model, introduced last month, the T1 modeal claims to ensure faster query processing compared to rival models, including DeepSeek’s R1.

The Tencent post stated that its T1 model sets new benchmarks for knowledge and reasoning capabilities, citing a performance chart highlighting T1’s superior results over DeepSeek R1 on select evaluations.

The launch follows heightened competition in China’s AI sector, driven by DeepSeek’s recent advancements in delivering high-performance models at lower costs.

Tencent had previously released a preview version of T1 through platforms like its AI assistant application, Yuanbao.

Tencent has also been doubling down on AI innovation, as evidenced by its plans to increase capital expenditure in 2025, building on significant AI investments throughout 2024.

By introducing the Tencent T1 reasoning model, Tencent itreaffirms its commitment to advancing AI technologies and maintaining a competitive edge in the rapidly evolving Chinese AI ecosystem.

Source of main image: Tencent media library