Tencent releases open-source Hy4 preview AI model

Tencent has released and open-sourced Hy4 preview, a new generation large language model (LLM) utilizing a Mixture-of-Experts (MoE) architecture. The model features 770 billion total parameters, with 49 billion parameters activated per token, and supports a context length of over 1 million tokens.

Designed for productivity, the model targets specialized fields including software engineering, scientific research, financial analysis, and game development. In internal blind tests involving 163 experts and 203 engineering tasks, Hy4 preview achieved an average score of 2.99 out of 4.00, slightly outperforming competitors such as Zhipu AI’s GLM 5.3 and Moonshot AI’s Kimi K3.

The model is released under the Apache License 2.0, making its weights available on platforms like Hugging Face and GitHub. Tencent plans to integrate the technology into its existing product ecosystem, including CodeBuddy and WorkBuddy. The company also noted that the model has participated in its own development process by assisting in the optimization of training methods and underlying operators.