analiza aiqbench

Z.ai has released its proprietary ZCode model and accompanying documentation following a security incident where user-submitted code was inadvertently processed and exposed. This move addresses significant concerns regarding data privacy and model safety, specifically highlighting risks similar to those found in other large language models where training data ingestion can lead to unintended information leakage. The open-sourcing of the model allows for independent auditing of these vulnerabilities within the Chinese AI ecosystem.