Theori Joins Anthropic's Security Program, Gaining Access to Both Major AI Platforms Alongside OpenAI
South Korean cybersecurity firm Theori has joined Anthropic's security-specialized artificial intelligence (AI) program. With this move, Theori has now secured access to advanced cybersecurity capabilities offered on a limited basis by both major generative AI companies—OpenAI and Anthropic.
Theori announced on the 16th that it has been selected as a participant in Anthropic's Cyber Verification Program (CVP). CVP is a program that grants access to advanced cybersecurity AI capabilities exclusively to security companies and experts who have passed a separate qualification review. Participants can leverage the latest AI models for high-difficulty security research, including malware analysis, vulnerability detection, and attack path validation.
The AI capabilities secured through the program will be deployed to enhance the analytical performance of Xint, Theori's autonomous penetration testing platform. Xint is a platform where AI independently identifies security vulnerabilities and verifies real-world exploitability, applying AI reasoning capabilities to source code and binary analysis as well as penetration testing in live service environments. The goal is to reduce false positives and rapidly identify high-risk vulnerabilities.
Theori specializes in "offensive security"—identifying and validating system vulnerabilities from an attacker's perspective. The company expects that joining Anthropic's CVP will further advance Xint's offensive security detection and attack path validation engine. In particular, combining the precise reasoning capabilities of AI agents with source code and binary analysis and live-service penetration testing is expected to significantly reduce false positives and accelerate the discovery of high-risk vulnerabilities.
Theori had previously received approval to participate in OpenAI's cyber defense program, Daybreak. With its addition to Anthropic's CVP, the company is now positioned to simultaneously leverage the security-specialized AI features that both companies offer on a restricted basis.
"As AI rapidly lowers the time and cost required for cyberattacks, defenders must also be able to actively leverage the most advanced AI technology available," said Park Se-jun, CEO of Theori. "We will continue to develop technologies that identify and validate real-world exploitable vulnerabilities faster and more accurately." He added that this approval represents "an important opportunity to apply world-class AI technology to a broader range of security research," and that Theori will support its customers in responding to threats before attackers can act.