This article first appeared on GuruFocus .

SentinelOne Inc. ( NYSE:S ), the endpoint and cloud security vendor, expanded its Wayfinder Frontier AI Services on Thursday. The offerings now run on OpenAI Daybreak models, delivered through the Daybreak Defense Network, and the rollout starts with GPT-5.6-Cyber.

The first of two new capabilities scans customer code repositories. It looks for OWASP-class flaws, implants, exposed secrets and supply-chain risk. Findings get mapped to MITRE ATT&CK. SentinelOne said its own offensive-security analysts validate every verdict before it reaches the customer. Compromise assessment evaluates telemetry against detection rules to surface posture gaps, including risky use of VPNs, proxies and remote-management tools, with results ranked by real-world exploitability. Both lean on the same underlying malware analysis workflow.

The model selection follows benchmarking work by SentinelLABS, the company's research arm, which found GPT-5.6-Cyber performed best in class on reverse engineering and analysis of military-grade malware. "Attackers are increasingly using AI to find and exploit weaknesses," said Chief Customer Officer Steve Stone.