Tencent unveils first flagship AI model with former OpenAI researcher at helm
The model is relatively small with only 295 billion parameters, bucking a recent trend of large models with trillions of parameters
The Shenzhen-based company said on Thursday that the new model, an open-source program called “Hy3 preview”, was its most powerful yet, on par with top Chinese models but still lagging flagship products from US leaders such as OpenAI and Google DeepMind.
Notably, the model is relatively small with only 295 billion parameters, bucking a recent trend of large models with trillions of parameters – the mathematical variables that encode a model’s “intelligence”, which are roughly proportional to the amount of computational power needed to train and serve the model.
The company’s previous flagship model, the HY 2.0 released in early December, had over 400 billion parameters.
The new model was developed to cater for real-world business scenarios, the company said, highlighting the collaboration between Tencent’s foundational model development team Hunyuan and its Yuanbao AI application team.
“By seamlessly aligning product-side requirements with underlying technology, we have successfully bridged the gap between model capability and user value,” it said, adding that the model was already deployed in Tencent’s flagship AI products including consumer app Yuanbao and coding assistant CodeBuddy.
Agentic capabilities were “one of the most significantly improved areas in this release”, with the company pointing to strong performance on several in-house benchmarks it had developed to assess compatibility with popular agentic AI tool OpenClaw.