Tencent expands computing resources for hy4preview model amid high demand

Tencent announced on August 31 that the launch of its Hy4preview model on the WorkBuddy platform on August 28 triggered significant user and developer engagement due to enhanced Agent capabilities. High concurrency demand led to waiting queues on the task queue, prompting Tencent to urgently expand the Hy4preview inference cluster and dynamically allocate resources to maintain service efficiency. To address the growing demand for high-performance inference computing power in complex Agent scenarios, Tencent extended the free usage rights of existing models like Hy3 on WorkBuddy until 23:59 on September 30, 2026. The surge in call volume highlights the increasing practical potential of AI Agents in real-world productivity scenarios, emphasizing the need for real-time reasoning and multi-step planning capabilities. Tencent stated that due to the limited scale of high-end computing power and concentrated concurrent calls during peak times, waiting situations may still occur during specific peak periods. The company emphasized the importance of flexible computing cluster expansion and multi-model coordination as core competitive advantages for AI infrastructure providers.