Tencent Hunyuan releases Hy4flash: 224B parameters, 1M context, decoding over 150 tps, single-task cost only 1/13 of Hy4preview
Tencent Hunyuan launched Hy4flash on Nov 16, its most cost-effective Hy model for daily use. 224B total params (64B Engram), 12B activated, multimodal, 1M context, much lower API pricing than Hy4. Available on Tencent Cloud TokenHub, WorkBuddy, Yuanbao, WeCom, QQ, ima, Marvis; coming to OpenRouter. Cost optimized end-to-end: hybrid attention, LatentMoE, Engram in pretraining; KV cache 1/24 of Hy3, 1/6 of Hy4.....