OpenAI connected Tuesday unveiled GPT-5.4-Cyber, a variant of its latest flagship exemplary fine-tuned specifically for defensive cybersecurity work, pursuing rival Anthropic’s announcement of frontier AI model Mythos.
Mythos, announced connected April 7, is being deployed arsenic portion of Anthropic’s “Project Glasswing”, a controlled inaugural nether which prime organisations are permitted to usage the unreleased Claude Mythos Preview exemplary for defensive cybersecurity purposes. It has recovered “thousands” of large vulnerabilities successful operating systems, web browsers and different software.

1 month ago
1
![The company is besides expanding its Trusted Access for Cyber programme [File] The company is besides expanding its Trusted Access for Cyber programme [File]](https://www.thehindu.com/theme/images/th-online/1x1_spacer.png)




