← Library · Frontier
Tencent Releases Hy3: A 295B Mixture-of-Experts Model
Tencent's Hy team has released Hy3, a 295 billion parameter Mixture-of-Experts (MoE) model. This architecture activates only 21 billion parameters per token, making it efficient for large-scale tasks. Hy3 is designed for reasoning, agentic workflows, and handling long-context tasks, with its weights available under the Apache License 2.0.
Why it matters
The release of Hy3 provides the open-source community with a powerful and efficient large language model, potentially accelerating advancements in agent-based AI and complex reasoning. Its MoE architecture offers a balance of scale and computational efficiency.
Learn one new AI thing every day.
Daily Deck sends you seven plain-English cards like this every morning. Free.
Start free