In Shanghai, Chinese tech company Tencent introduced the official version of its T1 reasoning model, aiming to strengthen its position in China's competitive artificial intelligence industry. The updated T1 model boasts quicker response times and improved capabilities for processing long text documents, as mentioned on Tencent's WeChat account. According to the company, T1 can maintain clear content logic and tidy text with a very low error rate.

The launch of T1 comes as competition intensifies in the Chinese AI sector, particularly with the recent release of models by DeepSeek that offer comparable or better performance than Western systems at lower prices. Tencent had previously unveiled a preview version of T1 on various platforms, including its Yuanbao AI assistant application.

The official version of T1 will leverage Tencent's Turbo S foundational language model, which was unveiled recently and is claimed to handle queries faster than DeepSeek's R1 model. A comparison chart shared in the announcement indicated Tencent's T1 model outperformed DeepSeek on certain knowledge and reasoning benchmarks.

Tencent has significantly increased its AI investments in recent months, with plans to ramp up capital expenditure through 2025 after substantial spending in 2024.