Models · MarkTechPost ·
Tencent Releases Hy3: An Open 295B Mixture-of-Experts (MoE) Model with 21B Active Parameters and 256K Context
Tencent’s Hy team released Hy3, an open-source 295B Mixture-of-Experts model with 21B active parameters per token and a 256K context window. The Apache 2.0 model is positioned for reasoning, agentic, and long-context tasks and is available to try on OpenRouter through July 21, 2026.