Tencent has released Hy4, its latest large language model, with cybersecurity training limited to defensive material in an effort to reduce its potential to aid cyberattacks.
Reuters reported that Tencent filtered out training content covering exploit development, reverse engineering and penetration testing.
The model’s general release adds vision understanding and computer-use capabilities, allowing it to work through graphical interfaces and command-line tools, according to Tencent’s Hy4 announcement.
Tencent said it plans to publish the model weights and a system card with fuller safety evaluations later. Hy4 is available through Tencent Cloud and OpenRouter and is being integrated into Tencent products.
Tencent said it recognizes the safety risks posed by advanced cyber and biological capabilities and is pursuing a more responsible approach to model training, open-sourcing and release. The company’s training restrictions are intended to limit offensive capabilities while retaining defensive cybersecurity applications.
If this article helped you, please consider supporting our work. Every small contribution keeps Abijita.com independent and running.
Tencent Hy4 Restricts Offensive Cybersecurity Training
