Models New
Tencent releases Hy4 flash, a cheaper model for everyday office tasks that it says costs a twentieth as much per task as Hy4
Published:
Tencent released Hy4 flash, the lowest-cost model in its Hy family, twelve days after the full Hy4. It has 224 billion total parameters, including 64 billion 'Engram' parameters, with 12 billion active, takes text and images and handles up to one million tokens of context.
List prices are about US$0.07 per million input tokens and US$0.281 per million output tokens. Tencent says that on WorkIndex, its own benchmark of slide, document and spreadsheet tasks, Hy4 flash costs one-thirteenth as much per task as Hy4 Preview, one-twentieth as much as Hy4 and one-eighth as much as DeepSeek-V4.1-Flash-1017, and that in a blind test by 85 internal WorkBuddy users its satisfaction came close to Claude Haiku 5.5.
The model is available in Tencent Cloud TokenHub, WorkBuddy, Yuanbao, WeCom, QQ and other Tencent products, free in WorkBuddy and at half price on TokenHub until 30 November. Tencent says it plans to release the weights and to add the model to OpenRouter, and recommends the full Hy4 for long software-engineering tasks.
Caveat. Cost and quality comparisons come from Tencent's internal benchmark and internal users, not independent tests. The weights are promised, not yet released.
Sources