Tencent releases Hy3 preview
295B-parameter open MoE model with 21B active, three reasoning modes and a 256K-token context window.
It is a 295B-parameter Mixture-of-Experts (MoE) model, with 21B active parameters, supporting three reasoning modes for latency/depth tradeoffs, and a 256K context window.
Official source
Hy3 preview: The First Step in Rebuilding the Hy modelhy.tencent.ai ยท published