Move over DeepSeek; there’s another Chinese-owned generative AI chatbot waiting to disrupt the artificial intelligence market — and this one claims that it’s even faster.
Tencent, one of the world’s biggest video game companies, has launched its new Hunyuan Turbo S model, with the promise of ‘instant reply’ responses to user prompts.
While it wasn’t so long ago that China’s ChatGPT challengers were struggling to keep pace with their US counterparts, the progress being made by the likes of Tencent, DeepSeek, and retailer Alibaba suggests that the country’s tech sector is now ready to lead the world in artificial intelligence.
‘Fast-Thinking’ AI
The Riot Games owner announced the launch of its newest GenAI model via a post on the official site for its Weixin (a.k.a. WeChat) messaging app. The model is available immediately.
Tencent calls Hunyuan Turbo S a ‘new generation fast-thinking’ model, that integrates long and short thinking chains to significantly improve ‘scientific reasoning ability’ and overall performance simultaneously.
This just in! View
the top business tech deals for 2025 👨💻
That, it says, means that Turbo S doesn’t rely on the ‘thinking before answering’ time required by DeepSeek R1 and its own Hunyuan T1 models.
“Different from slow thinking models such as Deepseek R1 and Hunyuan T1 that require “thinking before answering,” Hunyuan Turbo S can achieve “instant reply,” output answers more quickly, double the word pronunciation speed, and reduce the first word delay by 44%.” – Tencent
Tencent likens the speed of Turbo S to human intuition, which gives it “rapid response capabilities in general scenarios.” Adding that: “The combination and complementation of fast thinking and slow thinking can enable large models to solve problems more intelligently and efficiently.”
Benchmarking Success
In terms of architecture, Turbo S has adopted the Hybrid-Mamba-Transformer fusion mode – the first time, Tencent says, it has been successfully applied ‘losslessly’ to a very large model.
To demonstrate the model’s speed, the company lists benchmarking for Turbo S against DeepSeek-V3, OpenAI’s ChatGPT 4o, Anthropic’s Claude 3.5 Sonnet and Meta’s Llama 3.1 in areas including knowledge, reasoning, math and code.
Across the 17 sub-categories, it was the overall fastest in 10 (Claude 3.5 Sonnet performed next best, with five ‘wins’), beating ChatGPT 4o in 15 sub-categories and DeepSeek-V3 in 12.
“As the flagship model, Hunyuan Turbo S will become the core foundation of Tencent’s Hunyuan series of derivative models in the future, providing basic capabilities for derivative models such as reasoning, long texts, and codes.” – Tencent
Lowering Threshold for Large Model Applications
Tencent says that its innovative architecture also means that “the deployment cost of Hunyuan Turbo S has been greatly reduced, continuously lowering the threshold for large model applications.”
It was DeepSeek’s low cost, low resource model that helped catapult it to the top of the Apple App Store and Google Play Store in January. Chinese retail giant Alibaba since announced its own upgraded AI model that it claims outperforms DeepSeek and ChatGPT.
Due to its country of origin, however, it’s unlikely that the Riot Games owner will find cracking the international market plain sailing. Earlier in the year, the Tencent was designated a Chinese military company by the US Department of Defense, which will restrict US investment.
And DeepSeek has encountered its own issues, with Italy, Australia, South Korea and certain US states all moving to ban its use.
