Tencent touts latest AI model Hy4’s performance on independent safety assessments

What happened
Tencent worked with international AI safety organizations to conduct an independent risk assessment of its new open-weight AI model, Hy4. The evaluation focused on identifying potential dangers related to the model’s use. This marks the first time a major Chinese technology company has taken this approach. The assessment specifically examined risks such as AI-enabled cyberattacks and misuse in biological contexts.
Key details
- Tencent's new AI model is called Hy4.
- The model is open-weight, meaning its parameters are available for use.
- International AI safety organizations participated in the assessment.
- The evaluation included risks of cyberattacks and biological misuse.
- This is the first such independent assessment by a major Chinese tech firm.
Why it matters
Chinese technology companies, international safety groups, and users of open-weight AI models are directly affected. The independent assessment provides these groups with more information about potential risks, which can influence decisions about deploying or regulating the model. It also sets a precedent for other companies in China to follow similar safety evaluation practices.
