Open-weight AI models are catching up to the frontier. The safety gap remains. 

Researchers at SaferAI have found that Z.ai's open-weight GLM-5.2 model is nearing frontier AI capabilities, sparking concerns about the safety gap in open models. The GLM-5.2 model lacks key safety mitigations, despite its advanced capabilities, which could lead to unintended consequences. This development raises worries that powerful open models may outpace governance and safeguards, posing potential risks. The report highlights the need for improved safety measures in open-weight AI models to prevent potential harm.

✦ AI-generated summary