Tencent's Hy4 Flash delivers Hy4-level performance at one-thirteenth the cost

Tencent Hy4 Flash

Hy4 flash is Tencent's most cost-efficient model yet: 224B total parameters with only 12B active per token, native multimodality, and a 1M-token context window. On WorkIndex, an internal benchmark built from real WorkBuddy tasks, its average cost per task is 1/13 of Hy4 preview and 1/8 of DeepSeek-V4.1-Flash. In blind tests with 85 users, satisfaction neared Claude Haiku 5.5 while completion time was the shortest of any model evaluated.

In everyday office work, users care just as much about turnaround time and cost as they do about execution quality.

More from this day

2026-11-17