ModelScope Releases Hy3: 295B Total Param MoE Model with 256K Context

bdsqlsz · x · 2026-07-06

ModelScope has released the Hy3 model, featuring a Mixture-of-Experts (MoE) architecture with 295B total parameters and 21B activated parameters. Optimized specifically for agentic workflows, it supports a massive 256K context length and offers an FP8 quantized version to reduce inference costs.

Related event: Tencent Open-Sources Hunyuan Hy3 for Agentic Workloads(26 posts)→

Original post →

More from Models

Models channel →