Tencent has launched the official version of its T1 reasoning model, enhancing competition in China's AI sector with improved speed and accuracy.
 |
| Tencent debuts the full version of its T1 reasoning model, powered by the Turbo S language model, as competition intensifies in China’s AI industry. Image: Supplied |
SHENZHEN, China — March 23, 2025:
Tencent has launched the official version of its T1 reasoning model, stepping up competition in China’s artificial intelligence sector. The model, unveiled on Friday night, delivers faster response times and improved processing of long-form text, the company said in a post on its official WeChat account, citing a Reuters report.
T1 ensures “clear content logic and clean text,” with an “extremely low” hallucination rate, according to Tencent. The release follows heightened AI rivalry in China, particularly after DeepSeek introduced models that claim to rival or outperform Western systems at lower costs.
Tencent had previously rolled out a preview version of T1 on various platforms, including its AI assistant Yuanbao. The official version now runs on Tencent’s Turbo S foundational language model, which the company says offers faster query processing than DeepSeek’s R1 model.
A performance chart published alongside Tencent’s announcement indicated that T1 outperforms DeepSeek R1 on several knowledge and reasoning benchmarks.
Tencent has been increasing its AI investments, announcing on Thursday that it would boost capital expenditure in 2025 following aggressive spending in 2024. The company’s latest advancements reflect China’s broader push to lead in AI innovation amid growing global competition.