At the 2024 Tencent Global Digital Ecosystem Conference, Tencent's Vice President and Head of Cloud Business, Qiu Yuepeng, announced the official launch of the Tencent Hunyuan Turbo large model. This new generation of large models is designed based on the MoE (Mixture of Experts) architecture, achieving a doubling in inference efficiency with a 100% improvement compared to the previous generation, while significantly reducing inference costs by 50%. Additionally, Hunyuan Turbo also demonstrates outstanding performance in decoding efficiency, improving by 2.