LaunchAI ModelsNovember 16, 2026

Tencent launches Hy4 flash, a 224B-parameter low-cost model

Read original source →technode.com

Hy4 flash activates 12B parameters per token with a 1M-token context window; API list prices are $0.07/M input and $0.281/M output tokens. On Tencent's internal WorkIndex benchmark it cost one-twentieth as much per task as full Hy4, and in a 259-task blind test it scored near Claude Haiku 5.5 on user satisfaction.

3 sources

More stories today

Open the live feed