Tencent Launches Hunyuan Hy3 AI Model, Integrates Across Products
Tencent unveiled its new reasoning model, Hunyuan Hy3, on Monday, designed to integrate fast and slow thinking. The model has been integrated into several of the company's products and services.
Chinese technology giant Tencent launched its new large language model, Hunyuan Hy3, on Monday. The model utilizes a Mixture of Experts (MoE) architecture, featuring 295 billion total parameters with 21 billion activated parameters.
Huyuan Hy3 supports a context window of up to 256,000 tokens, enabling the processing of extensive text. The company stated the model integrates fast and slow thinking capabilities, enhancing its reasoning abilities.
The model has already been incorporated into several Tencent products and services, including WorkBuddy/CodeBuddy, Yuanbao, Marvis, and ima. Its API is now accessible through Tencent Cloud's TokenHub platform.
Tencent has set pricing for Hy3 at 1 RMB (approximately $0.15) per million input tokens and 4 RMB (approximately $0.59) per million output tokens. Cached input tokens are priced at 0.25 RMB (approximately $0.037) per million tokens.