Zscaler Partners with OpenAI to Integrate GPT‑5.4‑Cyber for Enhanced Security

Zscaler has partnered with OpenAI through the company’s Trusted Access for Cyber (TAC) program to strengthen cybersecurity across its Zero Trust Exchange platform, according to a company announcement. As part of this collaboration, Zscaler will integrate GPT‑5.4‑Cyber, a TAC‑enabled version of OpenAI’s latest model, into its secure software development lifecycle to identify and patch vulnerabilities more efficiently.

The GPT‑5.4‑Cyber model will assist Zscaler’s internal teams in validating threat models, performing secure code reviews, and conducting automated testing. It will also be used to transform large volumes of security data into actionable intelligence, helping prioritize real risks and accelerate remediation efforts. Zscaler plans to leverage the model defensively and offensively, including simulating adversarial attacks to improve system resilience.

Beyond development security, Zscaler is employing OpenAI models for its AI Red Teaming suite, which provides continuous testing, prompt hardening, and AI asset analysis. The models also support the company’s Red Canary Managed Detection and Response service, where AI agents assist analysts in investigating threats while maintaining high accuracy rates.

Through this partnership, Zscaler aims to combine OpenAI’s advanced models with its Zero Trust architecture to reduce attack surfaces and enhance protection for enterprise customers.