OpenAI’s latest vulnerability detection model, GPT-6 Cyber, is highly effective in identifying vulnerabilities in Apple products – it has found security holes in almost all Apple devices, including iPhone, Mac, iPad, Apple Watch, Apple TV and Vision Pro headsets, it said Forbes.

Image source: Mariia Berezovsky/unsplash.com
In recent months, OpenAI, along with its partner, cybersecurity firm Calif, has discovered approximately 35 vulnerabilities of varying severity in Apple hardware and software using GPT-6 Cyber.
Some of them could theoretically give attackers control of a Mac or iPhone via a malicious website. Researchers also used GPT-6 Cyber to create an exploit that can neutralize the protection provided by the Lockdown mode (strict security mode) in the iPhone. This feature intentionally limits the device’s capabilities to reduce the number of potential attack vectors.
Calif reported that Apple has already fixed many of the identified vulnerabilities immediately after receiving information about their discovery.
Calif CEO Thai Duong told Forbes that GPT-6 Cyber is significantly faster than its predecessors in the speed of creating exploits, which may be useful for quickly eliminating vulnerabilities, but if attackers gain access to it, it can become a serious threat. With its help, the company discovered a bug in Java memory management mechanisms called Java4Shell. Calif also discovered a vulnerability in PHP in August, but it used the GPT-5.6 Sol AI model.
If you notice an error, select it with the mouse and press CTRL+ENTER.


