IDC Releases First Large Model Security Assessment Report Featuring 360

The International Data Corporation recently released a report evaluating large model application firewall technology in China. Among fifteen major domestic vendors assessed, the 360 Digital Security Group achieved the highest ratings across five dimensions, including content security, data protection, and technical innovation, ultimately securing the top position in overall rankings. The report indicates that the artificial intelligence super cycle has arrived, with significant growth projected for the generative market and enterprise intelligent agents. As these models integrate into core business operations, traditional security systems struggle to address emerging risks like prompt attacks and data leaks, necessitating new AI security infrastructure. To address these challenges, 360 utilizes an approach centered on using models to govern models and promoting defense through testing. This system employs multiple professional security models to identify risks and constrain behaviors, creating a closed loop that covers risk identification before deployment, real-time control during operation, and subsequent optimization. The 360 system extends protection from content to behavioral safety by implementing a three-dimensional security logic. This framework manages risks related to intelligent agent tool calls, high-risk instructions, and sensitive data access. By controlling these specific areas, the company establishes necessary security boundaries for artificial intelligence tasks during their execution.