By Yelin Mo and Andrew Silver
SHANGHAI, March 22 (Reuters) - Chinese tech giant Tencent on Friday night launched the official version of its T1 reasoning model, stepping up competition in China's increasingly crowded artificial intelligence sector.
The upgraded T1 model offers faster response times and enhanced capabilities for processing extended text documents, the company said in a post on its official WeChat account.
T1 can "keep the content logic clear and the text neat and clean", the post said, while the hallucination rate is "extremely low".
The launch comes amid heightened rivalry in China's AI landscape following DeepSeek's introduction of models that offer comparable or superior performance to Western systems at substantially lower costs.
Tencent 0700.HK had previously released a preview version of T1 through platforms, including its AI assistant application Yuanbao.
The official version will be powered by Tencent's Turbo S foundational language model, unveiled late last month, which the company claims processes queries faster than rival DeepSeek's R1 model.
A chart published in the post comparing the T1 model to DeepSeek R1 showed Tencent's outperformed on some knowledge and reasoning benchmarks.
Tencent has ramped up its AI investments in recent months. On Thursday, the company announced plans to increase capital expenditure in 2025, following already aggressive AI spending throughout 2024.
(Reporting by Yelin Mo and Andrew SilverEditing by Shri Navaratnam)
((andrew.silver@thomsonreuters.com;))
免責聲明：投資有風險，本文並非投資建議，以上內容不應被視為任何金融產品的購買或出售要約、建議或邀請，作者或其他用戶的任何相關討論、評論或帖子也不應被視為此類內容。本文僅供一般參考，不考慮您的個人投資目標、財務狀況或需求。TTM對信息的準確性和完整性不承擔任何責任或保證，投資者應自行研究並在投資前尋求專業建議。